Abstract

We compared 156 human immunodeficiency virus (HIV)–infected patients who had tuberculosis with control populations of similar size. Of 111 patients with HIV infection and tuberculosis who received highly active antiretroviral therapy (HAART) and therapy for tuberculosis concurrently, 92 (83%) achieved or maintained virus loads of <50 copies/mL, and 99 (89%) achieved or maintained a ⩾2 log10 reduction in virus load after 6 months. Virological response and changes in CD4 cell count were equivalent to those in 111 matched HIV-infected subjects without tuberculosis starting HAART. Tuberculosis recurrence rates were similar to those found in an HIV-uninfected population of 156 subjects (3% and 1%, respectively). Treatment for HIV and tuberculosis does not compromise outcomes for either disease

The incidence of tuberculosis and HIV infection continue to increase, with devastating global consequences [1]. Although treatment of tuberculosis takes priority initially, effective treatment with highly active antiretroviral therapy (HAART) improves long-term outcome in HIV-infected persons with advanced immunosuppression [2]. Combining therapy for tuberculosis and HIV introduces potential practical problems, including overlapping toxicities, inflammatory reactions to both antituberculosis therapy and HAART, high pill burden, drug-drug interactions, and variable drug absorption [3]. All of these problems may have a negative impact on the effectiveness of HAART, which, in turn, may lead to virological failure. We hypothesized that specialist care provided by a team experienced in treating both tuberculosis and HIV could overcome these potential difficulties and that response to HAART and antituberculosis therapy should be as robust as that seen in other populations requiring these treatments

MethodsRetrospective data collection was performed at the North Middlesex, Royal Free, and University College Hospitals in London. We identified consecutive, unselected, adult HIV-infected persons with tuberculosis presenting between February 1997 and November 2003. A retrospective case-note review was performed to identify patients who received HAART and antituberculosis therapy concomitantly. Further data for these patients were extracted from the case notes. Tuberculosis was diagnosed if a patient had a positive result of culture for Mycobacterium tuberculosis; had negative results of culture but positive results of nucleic acid amplification assay (TB Strand Displacement Amplification assay; Becton Dickinson), with clinicoradiological features and response to treatment consistent with tuberculosis; or had histological findings and response to treatment consistent with tuberculosis. HAART was defined as the use of at least 3 antiretroviral drugs in combination. Data were recorded on a predesigned form and entered into a database

The virological responses of patents with tuberculosis to HIV therapy were compared with those observed in a control group of 111 antiretroviral-naive persons without tuberculosis who started HAART during the study period. Because the demographic distribution of our general HIV-infected population is different from that of the HIV-infected population with tuberculosis, which might affect observed response rates, cases and controls were matched 1:1 for sex, ethnicity, baseline CD4 cell count (to within 50 cells/μL), and type of HAART regimen. Cases and controls were not matched for year of starting HAART, although all did so between February 1997 and November 2003, along with the patients with tuberculosis. Because several HIV-infected patients without tuberculosis had virus load measurements made using assays with lower limits of detection of 400 copies/mL, we compared the proportion of persons with and without tuberculosis who attained a virus load of <400 copies/mL within the first 6 months of HAART. Comparisons of virus load response between the 2 groups were made using McNemar’s test to account for matching between individuals, and comparisons of the CD4 cell response were made using the signed&amp;rank test [4]

Outcomes of antituberculosis therapy were compared with those for a second control population of 156 consecutive, unselected, unmatched, HIV antibody–negative patients treated for tuberculosis at these centers (where HIV antibody testing of all tuberculosis patients is the standard of care) during the study period. Data for patients included in this comparison group were collected retrospectively using the same criteria as for the group with HIV and tuberculosis, and data collection was continued until identical numbers had been analyzed in both groups

HAART and antituberculosis therapy were prescribed at the discretion of individual physicians using the then-available treatment protocols. In all cases, initial antituberculosis therapy was rifamycin-based and adjusted according to the availability of susceptibility testing. Rifampicin (600 mg if >50 kg body weight; 450 mg if <50 kg body weight) was prescribed except with single or “boosted” protease inhibitor–containing regimens, when rifabutin was given at a dose of either 150 mg daily or 150 mg 3 times per week, respectively. Standard practice was to prescribe all drugs as daily self-administered treatment

ResultsDuring the study period, 156 HIV-infected patients treated for tuberculosis were identified. One hundred eleven (71%) received HAART and antituberculosis therapy concurrently. The median age of these 111 patients was 35 years (range, 24–73 years), 52% were men, 77% were black-African, and 85% had positive results of culture for M. tuberculosis. The median CD4 cell count and virus load at diagnosis of tuberculosis were 77 cells/μL (range, 0–720 cells/μL) and 5.1 log10 copies/mL (range, 1.7–5.9 log10 copies/mL), respectively. When tuberculosis was diagnosed, 29 (26%) of 111 were already receiving HAART, which had been started a median of 18 weeks (range, 1–85 weeks) before antituberculosis therapy (18 of 29 patients commenced HAART 1–23 weeks before antituberculosis therapy, and 11 of 29 patients commenced HAART 37–85 weeks before antituberculosis therapy). In 15 (52%) of 29 patients, a virus load of <50 copies/mL had been achieved before antituberculosis therapy. By the end of antituberculosis therapy, the virus loads of all 15 remained undetectable, and 11 others had achieved this result. Thus 26 (90%) of 29 had virus loads of <50 copies/mL

For the remaining 82 patients, HAART was commenced a median of 2 months (range, 0–8 months) after starting antituberculosis therapy. Of these 82 patients, 68 (83%) had a virus load of <50 copies/mL 6 months after initiation of HAART, and 76 (93%) had a ⩾2 log10 reduction in virus load. No difference in virological response was seen according to whether HAART was started within the first 2 months of initiation of antituberculosis therapy (32 [39%] of 82 patients) or >2 months after initiation of antituberculosis therapy (50 [61%] of 82 patients). Therefore, among all who received concurrent HAART and antituberculosis therapy, 92 (83%) of 111 patients achieved or maintained a virus load of <50 copies/mL, and 99 (89%) of 111 patients achieved a ⩾2 log10 reduction in virus load within 6 months of starting concurrent therapy

Among the 82 patients who initiated HAART following diagnosis of tuberculosis, 71 (87%) had an increase in CD4 cell count 6 months after starting HAART (median, 108 cells/μL; range, 10–460 cells/μL). Eleven (13%) patients had no increase in CD4 cell count (median, −20 cells/μL; range, −460 to 0 cells/μL). Three of these 11 patients had achieved a virus load of <50 copies/mL by this time

Five HAART regimens, involving 31 different combinations of drugs, were prescribed (table 1). No differences according to the type of HAART regimen were seen in the numbers of patients achieving a virus load of <50 copies/mL or a ⩾2 log10 reduction (P=.63, P=.50, respectively), although numbers in some groups were small. Baseline demographics, tuberculosis culture results, CD4 cell count, year of starting HAART, or treatment center did not differ between those who did and those who did not achieve an undetectable virus load (data not shown)

Table 1

Comparison of virological responses according to regimen of highly active antiretroviral therapy (HAART)

Table 1

Comparison of virological responses according to regimen of highly active antiretroviral therapy (HAART)

The patients with tuberculosis and HIV infection who achieved a virus load of <400 copies/mL within 6 months of starting HAART were compared with 111 HIV-infected persons without tuberculosis. Baseline features of the matched control group were as follows: a median age of 35 years (range, 20–74 years), a median baseline CD4 cell count of 102 cells/μL (range, 1–736 cells/μL), and a median baseline virus load of 5.3 log10 copies/mL (range, 3.0–6.3 log10 copies/mL). No difference in virological response was seen between the patients with HIV and tuberculosis and the control group. Fourteen (13%) of 111 patients with HIV infection and tuberculosis failed to achieve a virus load of <400 copies/mL within 6 months of starting HAART, compared with 13 (12%) of 111 persons without tuberculosis (P=.84). When only those 82 persons who started HAART after starting antituberculosis therapy were analyzed, we observed that 71 (87%) of 82 patients with tuberculosis achieved a virus load of <400 copies/mL within 6 months of starting HAART, compared with 75 (91%) of 82 persons without tuberculosis (P=.35, McNemar’s test). Similar changes in CD4 cell counts from baseline to 6 months were observed between both groups, with a median change of +97 cells/μL (range, −160 to +460 cells/μL) in the group with tuberculosis and of +89 cells/μL (range, −62 to +647 cells/μL) in the group without tuberculosis (P=.73, signed&amp;rank test)

Tuberculosis outcomes in the population with HIV infection and tuberculosis were compared with those in an unselected group of consecutive HIV-uninfected patients with tuberculosis (table 2). Nine (6%) of 156 received less than their planned course of antituberculosis therapy, compared with 6 (4%) of 156 of HIV-uninfected patients with tuberculosis. Follow-up data were available for 151 (97%) of 156 patients with HIV infection and tuberculosis after completion of treatment and for 150 (96%) of 156 HIV-uninfected patients. Tuberculosis recurred in 4 (3%) of 151 patients with HIV infection and tuberculosis during a median follow-up of 38 months (range, 8–90 months); 76% were followed for ⩾2 years. By comparison, tuberculosis recurred in 2 (1%) of 150 HIV-uninfected patients during a median follow-up of 20 months (range, 2–60 months); 45% were followed for ⩾2 years

Table 2

Clinical and demographic data for HIV-infected persons with tuberculosis and a control group of HIV-uninfected patients with tuberculosis

Table 2

Clinical and demographic data for HIV-infected persons with tuberculosis and a control group of HIV-uninfected patients with tuberculosis

DiscussionThese data suggest that persons coinfected with HIV and M. tuberculosis may be treated, with good outcomes for both conditions. We observed robust responses to HAART in most subjects who received this concurrently with antituberculosis therapy. HAART could be successfully combined with antituberculosis therapy regardless of it being commenced before or after the diagnosis of tuberculosis had been made. Furthermore, no difference in virological response was observed whether HAART was commenced within 2 months of diagnosis of tuberculosis or after this time. The virological response in the coinfected population was similar to that in a matched group with HIV infection alone starting HAART, whereas tuberculosis outcomes were similar to those in a group of consecutive HIV-uninfected patients with tuberculosis

These data concur with previously published work. Hung et al. [5] compared HIV-infected persons with and without tuberculosis starting HAART; in their cohort, 44% of patients with tuberculosis and 47% of those without tuberculosis achieved undetectable virus loads (<400 copies/mL) after receiving HAART for 1 month. However, among patients receiving HAART for at least 4 months, virological failure occurred in 38% with tuberculosis and 26% without tuberculosis. The treatment responses seen in our cohort, including patients receiving HAART at the time of diagnosis of tuberculosis and HAART-naive patients, were sustained after 6 months of therapy

Many of our patients are from sub-Saharan Africa, and our observations are in line with results from a South African study showing that, among 21 patients with tuberculosis and HIV, 94% had undetectable virus loads (<40 copies/mL) after 3 months of HAART [6]. The authors implied that the good outcome arose from use of directly observed therapy, although our data suggest that an equivalent outcome may be achieved with unsupervised, patient self-administered treatment

Previous work has focused largely on persons starting HAART after initiating antituberculosis therapy. However, a significant proportion of HIV-infected patients have been shown to present with tuberculosis while already undergoing HAART [7–9]. Thus, our observation that continuation of HAART was generally associated with virological suppression, regardless of whether this had been achieved prior to diagnosis of tuberculosis, is a novel finding

In our study, HAART was initiated at the discretion of individual physicians. This resulted in 31 different combinations of 5 specific treatment regimens and initiation of HAART at a wide range of times. We suggest that this diversity arose from the desire to tailor treatment to individual patients’ needs, which resulted in a positive outcome. This strategy may be difficult to follow in situations in which limited HAART regimens are available and less flexibility exists with regard to treatment initiation. Reassuringly, the likelihood of virological success appeared to be unrelated to either the specific HAART regimen used or its time of initiation

We found similar tuberculosis outcomes for both HIV-infected and HIV-uninfected patients with tuberculosis in terms of treatment completion and recurrence of tuberculosis. Indeed, the similar recurrence rates were achieved despite a substantially shorter follow-up for HIV-uninfected patients with tuberculosis than for the HIV-infected group. This observation concurs with that of Murray et al. [10], who described an African population not undergoing HAART, in contrast to our largely African population with access to antiretrovirals

Our study is limited by its retrospective design and by the difference in ethnicity between the group with HIV and tuberculosis and the HIV-uninfected control population. This reflects the demographics of patients with tuberculosis seen in London. However, we believe that all patients received an equivalent standard of care, because care was delivered by the same physicians. The diversity of times of initiation of HAART also prevents our study from providing insight as to when it is optimal to commence HAART in coinfection with HIV and M. tuberculosis. A further limitation is that direct extrapolation of our findings to resource-poor environments with a high burden of tuberculosis may not be possible, despite our patient population being largely from such areas

In conclusion, these data demonstrate that successful treatment of HIV infection and tuberculosis is possible and that achieving virological responses as robust as those seen in HIV-infected persons without tuberculosis is feasible

References

1
Corbett
EL
Watt
CJ
Walker
N
, et al.  . 
The growing burden of tuberculosis
Arch Intern Med
 , 
2003
, vol. 
163
 (pg. 
1009
-
21
)
2
Dheda
K
Lampe
F
Johnson
MA
Lipman
MCI
Outcome of HIV-associated tuberculosis in the era of highly active antiretroviral therapy
J Infect Dis
 , 
2004
, vol. 
190
 (pg. 
1670
-
6
)
3
Burman
WJ
Issues in the management of HIV-related tuberculosis
Clin Chest Med
 , 
2005
, vol. 
26
 (pg. 
283
-
94
)(pg. 
vi
-
vii
)
4
A Petrie
C Sabin
Medical statistics at a glance
 , 
2005
2nd ed
London
Blackwell
5
Hung
CC
Chen
MY
Chin-Fu
H
, et al.  . 
Improved outcomes of HIV-1–infected adults with tuberculosis in the era of highly active antiretroviral therapy
AIDS
 , 
2003
, vol. 
17
 (pg. 
2615
-
22
)
6
Cassol
E
Page
T
Mosam
A
, et al.  . 
Therapeutic response of HIV-1 subtype C in African patients coinfected with either Mycobacterium tuberculosis or human herpesvirus-8
J Infect Dis
 , 
2005
, vol. 
191
 (pg. 
324
-
32
)
7
Girardi
E
Palmieri
F
Cingolani
A
, et al.  . 
Changing clinical presentation and survival in HIV-associated tuberculosis after highly active antiretroviral therapy
J Acquir Immune Defic Syndr
 , 
2001
, vol. 
26
 (pg. 
326
-
31
)
8
Breen
RAM
Smith
CJ
Cropley
I
, et al.  . 
Does immune reconstitution syndrome promote active tuberculosis in patients receiving highly active antiretroviral therapy
AIDS
 , 
2005
, vol. 
19
 (pg. 
1201
-
6
)
9
John
L
Baalwa
J
Kalimugogo
P
, et al.  . 
Response to “Does immune reconstitution promote active tuberculosis in patients receiving highly active antiretroviral therapy?
AIDS
 , 
2005
, vol. 
19
 (pg. 
2049
-
50
)
10
Murray
J
Sonnenberg
P
Shearer
S
, et al.  . 
Human immunodeficiency virus and the outcome of treatment for new and recurrent pulmonary tuberculosis in African patients
Am J Respir Crit Care Med
 , 
1999
, vol. 
159
 (pg. 
733
-
40
)
Potential conflicts of interest: none reported