Increased Mortality and Tuberculosis Treatment Failure Rate among Human Immunodeficiency Virus (HIV) Seropositive Compared with HIV Seronegative Patients with Pulmonary Tuberculosis Treated with "Standard" Chemotherapy in Kinshasa, laire1- 3
JOSEPH H. PERRIENS, ROBERT L. COLEBUNDERS, CIBANGU KARAHUNGA, JEAN-CLAUDE WILLAME, JACQUES JEUGMANS, MULAMBA KABOTO, YIADIUL MUKADI, PAUL PAUWELS, ROBERT W. RYDER, JACQUES PRIGNOT, and PETER PlOT
Introduction
Many reports from both industrialized and developing countries have demonstrated a strong association between tuberculosis and human immunodeficiency virus (HIV) infection (1-12). Patients with inactive tuberculosis infection who subsequently acquire HIV infection have been shown to be at greatly increased risk of developing active tuberculosis (13). Consequently, there is a general consensus that the HIV epidemic will greatly complicate tuberculosis control (14). Antituberculosis chemotherapy is one of the mainstays of tuberculosis control, and chemotherapy including rifampin is able to achieve similar antimicrobial efficacy in patients with and without HIV infection (3, 5). There are, however,conflicting data on the efficacy of "standard" long-duration antituberculosis therapy regimens, including streptomycin, isoniazid, and thiacetazone, which are still in widespread use in many developingcountries (9, 10). The aim of the present study was to assess the treatment outcomes of HIV seropositive and HIV seronegative patients with pulmonary tuberculosis treated according to a streptomycin-based treatment policy. Methods The study wasconducted in Kinshasa, the capital city of zaire, where 70070 of the adult population is purified protein derivative positive, where the incidence of pulmonary tuberculosis is estimated to be 3/1,000 inhabitants/year, and where 3 to 8% of the general adult population is estimated to be HIV seropositive (15, 16). 750
SUMMARY Toevaluate their treatment outcomes 170human Immunodeficiency virus (HIV) seropositive and 597 HIV seronegative patients with active pulmonary tuberculosis (TB) treated for 1 yr with "stsndard" chemotherapy, Including streptomycin, Isoniazid, and, In most cases, thiacetazone, were traced at completion of therapy. All 582 survivors were Invited for reevaluation, and 385 patients, of whom 82 (21.3%) were HIV seropositive, were evaluated. Of those, 325 consenting patients, of whom fJ1 (20.6%) were HIV seropositive, were followed for 12 months. One year after TB had been diagnosed 47 (31.3%) of the 150 HIV seroposltlva and 22 (4.4%) of the 501 HIV seronegatlva patients traced had died (p 10-'). During the subsequent year the mortsllty of fJ1 HIV seropositive patients (26.31100 patlent-years) was higher than that of the 303 HIV seronegative patients (2.21100 patlents-years, p = 10-'). HIV seropositive patients hed a higher overall TB therapy failure rate 24 months after the diagnosis of TB than did HIV seronegative patients (21.11100 patlent-years versus 8.11100 patient-years, p 0.002), mainly because their relapse rate of pulmonary TB (18.11100 patient-years) wes higher than that of HIV seronegative patients (6.01100 patient-years, p 0.03). Given their higher relapse rate after 1 yr of "stendard" chemotherapy, the public health Impact of routine maintenance therapy In HIV seropositive patients with pulmonary TB who complete such therapy should be assessed In comparison to the Introduction of rlfamplcln-based short-eourse antituberculosis chemotherapy In developing countries. AM REV RESPIR DIS 1991; 144:750-755
=
=
Whenever feasible patients with tuberculosis are treated as outpatients in primary health care centers. "Standard" chemotherapy for pulmonary tuberculosis in adults weighing more than 50 kg consists of daily intramuscularly administered streptomycin (l g) and orally administered thiacetazone (150 mg) plus isoniazid (300 mg) for 2 months, followed by daily orally administered thiacetazone (150 mg) plus isoniazid (300 mg) for 10 months. If thiacetazone is unavailable, if the patient refuses to use it, or if side effects to thiacetazone occur, daily streptomycin (1 g) and isoniazid (300 mg) for 2 months, followed by isoniazid (900 mg) twice a week plus streptomycin (1 g) for 10 months is given instead (17). Therapy failures and relapses of pulmonary tuberculosis occurring within 12 months of completion of chemotherapy receivea retreatment scheme that consists of 2 months daily rifampicin (600 mg), ethambutol (1,200 mg), isoniazid (600 mg), and pyrazinamide (2,000
=
mg), followed by 4 months rifampicin (600 mg) and isoniazid (900 mg) twice a week (17). All study patients were treated according to this treatment policy. None of the HIV
(Receivedin originalform August 13, 1990and in revised form April 16, 1991) 1 From Projet Sida, the Bureau National de la Thberculose, the Centre de Depistage de la Thberculose, Kinshasa, Zaire; the Institute of Tropical Medicine, Antwerp, the Belgo-Zairian Medical Cooperation, Brussels, and the Universite Catholique de Louvain, Mont Godinne, Belgium; the Center for Infectious Disease, Centers for Disease Control, Atlanta, Georgia. 2 Supported by the Belgo-Zairian Medical Cooperation and by the Banque Commerciale du zaire. 3 Correspondence and requests for reprints should be addressed to J. Perriens, Institute of 'Iropical Medicine, Department of Microbiology, Nationalestraat 155, B-2000 Antwerp, Belgium.
PULMONARY TUBERCULOSIS AND HUMAN IMMUNODEFICIENCY VIRUS INFECTION
seropositive patients was treated with antiretroviral drugs.
Study Design The study was conducted in four phases. (1) Patient identification: HIVserosurveys among patients with tuberculosis. In 1987, t '.voHIV serosurveys were performed in Kin-
shasa among inpatients and outpatients with tuberculosis (9, 12). The HIV serosurvey among inpatients was done in May 1987 in a sanatorium. Outpatients with smear-positive pulmonary tuberculosis were studied in Kinshasa's biggest tuberculosis diagnosis center, the Centre de Depistage de la Tuberculose, between January and April 1987.
(2) Chart review at the end ofantituberculosis chemotherapy. One year later, between January 1988 and May 1988, when the patients identified by the HIV serosurveys had completed antituberculosis therapy, their treatment charts were reviewed, and mortality during treatment was recorded.
(3) Patient evaluation at the end of antituberculosis therapy. Surviving patients, whose treatment charts were found, and who could be physically located, were invited for reevaluation (between January and May 1988). Patients who showed up for evaluation were interviewed concerning their symptoms, and a physical examination, repeat HIV serology, sputum microscopy, and sputum culture were performed. (4) Follow-up after completion of antituberculosis therapy. Consenting patients,
evaluated at the end of chemotherapy, were enrolled in a longitudinal study, which lasted from January 1988to May 1989. In this study, history, physical examination, and sputum microscopy were repeated bimonthly. Resources to do sputum cultures were not available in this phase of the study. Mortality was recorded.
Diagnostic Procedures Sputum microscopy was done on freshly expectorated sputum, induced by nebulization of hypertonic 3070 saline if sputum could not be spontaneously produced, homogenized, and decontaminated using the sodium laurylsulfate method (18), and centrifuged for 30 min at 3,000 x g. Ziehl-Neelsen-stained smears of the sediment were examined under light microscopy by one of us (P. Pauwels). Sputum cultures were performed in duplicate on Lcwenstein-Jensen medium (Bacto Loewenstein Medium Base; Difco Laboratories, Detroit, MI) and Ogawa medium (19). The mycobacterial strains were identified using nitrate production, niacin production, and thiophene carboxylic acid hydrazide susceptibility tests as described by Jenkins and coworkers (20). Susceptibility tests for streptomycin, isoniazid, thiacetazone, ethambutol, and rifampicin were done on LOwenstein media using the absolute concentration method (21). HIV-serology was done in duplicate using an ELISA test (Wellcozyme; Wellcome Diagnostics, Dart ford, UK, or Organon Vironostica, Organon Teknika, Oss, The
Netherlands). Repeatedly reactive samples by ELISA were confirmed by indirect immunofluorescence in the HIV serosurvey among outpatients, and by Western blot (Dupont de Nemours, Wilmington, DE) in the serosurvey among inpatients and at the end of chemotherapy. A Western blot was considered positive if antibodies against p24 and gp41 and/or gp120/160 were present.
Definitions Patients with sputum smears positive for acidfast bacilli or sputum cultures positive for Mycobacterium tuberculosis were considered to have active pulmonary tuberculosis. Smear-positive patients and patients with sputum cultures positive for M tuberculosis 12 months after starting chemotherapy were considered therapy failures. Patients who at reevaluation after chemotherapy had sputum smears and sputum cultures negative for mycobacteria, and who later had sputum smears positive for acid-fast bacilli, were considered to have a relapse of pulmonary tuberculosis. Patients who initially (at the time the HIV serosurveys were done) had started treatment for a therapy failure or a relapse will be called "retreatments" in this report. Other patients will be called "new cases." The overall tuberculosis therapy failure rate was considered to be the therapy failure rate plus the relapse rate. A patient was considered compliant with therapy if he had completed at least 11 of the 12 months therapy prescribed, at least 56 of the 60 streptomycin injections prescribed during the first 2 months of chemotherapy, and if he had never interrupted therapy for more than 2 consecutive weeks.
Statistical Analysis The chi-square test (with Yates correction in 2 x 2 tables), Fischer's exact test, the t test, and the Mann-Whitney test were used as appropriate. The significance level used was p < 0.05.
Results
Patient Population The patient flow through the study is outlined in table 1,and characteristics of patients in/excluded in subsequent phases of the study are outlined in table 2. A total of 767 patients with microbiologically confirmed pulmonary tuberculosis was examined in the HIV serosurveys, of whom 170 (22.20/0) were HIV seropositive. HIV seroprevalence was lower among outpatients (124/621, 20%) than among inpatients (46/146, 31.5%; p < 0.01), and lower among new cases (124/642, 19.3%) than among retreatments (46/125,36.8%; p < 0.01). Female patients were more often HIV seropositive (87/319,27.3%) than were male patients (83/448, 18.5%; p = 0.005). HIV
751
seropositive patients were on average somewhat older (mean, 31.8 yr; SO, 8.4 yr) than were HIV seronegative patients (mean, 29.8 yr; SO, 10.7 yr; p = 0.001). Demographic characteristics of inpatients and outpatients were similar. The characteristics of patients included in the chart review after therapy, of patients who were subsequently reevaluated, and of patients who werethereafter followed, were similar to the ones of patients included in the HIV serosurveys for age and sex, the distribution of newcases and retreatments, and the distribution of inpatients and outpatients. Of the patients known to be alive after the chart review was completed, a higher proportion of HIV seropositive (82/103, 79.6%) than of HIV seronegative patients (303/479,63.3%) was subsequently reevaluated (p < 0.002). This was a consequence of more intensive follow-up efforts aimed at the HIV seropositive patients. Consequently, the proportion of HIV seropositive patients remained constant throughout the study despite the higher mortality of HIV seropositive patients during therapy. Nine patients, seven seroconversion and two initially HIV seropositive who were HIV seronegative on reevaluation, were excluded from the study, as the original samples from HIV serosurveys were not available for retesting.
Chart Review: Mortality during the Year after the Start of Antituberculosis Chemotherapy The one-year mortality among HIV seropositivepatients was 31.30/0 (47/150), significantly higher than the 22/501 (4.4%) among HIV seronegativepatients (relative risk, 7.0; 95% CI, 4.0 to 12.6; p < 10-6 ) (table 3). The mortality of HIV seropositive new cases and retreatments from both serosurveys was similar, but among HIV seronegative patients, retreatments had a higher mortality rate (10.5%) than among new cases (3.6%; p < 0.05). Patient Evaluation at the End of Chemotherapy Efficacy ofantituberculosis therapy. At the end of chemotherapy there were 13 (15.9%) therapy failures among 82 HIV seropositive patients and 40 (13.2%) among 303 HIV seronegative patients (p = 0.65) (table 4). Of the 53 therapy failures, diagnosis in two (3.7%) was made by sputum smear only, eight (15.1%) were smear positive and culture positive.and 43 (81.1 %) wereculture positive only.
PERRI~NS, COLEBUNDERS, KARAHUNGA, ET AL.
752 TABLE 1
PATIENT FLOW BY HUMAN IMMUNODEFICIENCY VIRUS (HIV) SEROLOGIC STATUS HIV (-)
HIV (+) Patient identification: HIV serosurveys (January-May 1987)
46 Inpatients 124 Outpatients
100 Inpatients 497 Outpatients
170 Total
597 Total
Chart not found: 20 Chart found: 150
Chart review at the end of chemotherapy (January-May 1988)
Chart not found: 96 Chart found: 501
IDead: 22
IDead: 47 Alive: 103
Alive: 479 Excluded: 7
Excluded: 2
Patient evaluation at end of chemotherapy (January-May 1988)
Not evaluated: 19 IEvaluated: 82
INot evaluated:
ITB not cured:
ITB not cured: 40 TB cured: 263
13
TB cured: 69 Longitudinal study after end of therapy (January 1988-May 1989)
Follow up on relapse
Refused FU: 11 Included in FU: 58 Relapsed: 8
No relapse: 50 (including 26 with FU < 12 months)
169
Evaluated: 303
Follow up on mortality
Refused FU: 15 Included in FU: 67
Follow up on relapse
Follow up on mortality
Refused FU: 38 Included in FU: 225
Refused FU: 45 Included in FU: 258
Died: 15
Died: 5
Relapsed: 12
Alive: 52 (including 29 with FU < 12 months)
Alive: 253 (including 53 with FU < 12 months)
No relapse: 213 (inclUding 49 with FU < 12 months)
Definition of abbreviations: FU = follow up; TB = tu berculosis.
The only mycobacterium isolated from the 51patients with positive sputum cultures at the end of chemotherapy was M tuberculosis. Susceptibility tests were done on 50 isolates. Only three were sensitiveto all fivedrugs tested. Isolates from HIV seropositive patients were resistant
to fewer drugs (on average, 1.5 drugs; 950/0 CI, 0.9 to 2.1) than isolates from HIV seronegative patients (on average, 2.5 drugs; 95% CI, 2.2 to 2.8; p = 0.003). Resistance to individual antimicrobial agents as 48% for streptomycin, 92% for isoniazid, 68% for thiacetazone, 6% for
rifampicin, and 12% for ethambutol. As susceptibility tests were not performed at the start of therapy, the causal relationship of resistance with the occurrence of therapy failures could not be analyzed. Noncompliance with chemotherapy was more frequent among HIV seroposi-
TABLE 2 CHARACTERISTICS OF THE PATIENT POPULATION
Patient identification Chart review at the end of chemotherapy Included Lost Reevaluation at the end of chemotherapy Included Not included Follow-up on mortality after chemotherapy Included Not included
HIV(+) Patients
Mean ± SO Age at Diagnosis
Male Patients
Ex-inpatients
New Cases
(%)
(n)
(%)
(n)
(%)
448
58.4
146
19.0
642
83.6
29.6 ± 10.7 30.7 ± 10.7
370 78
56.8 67.2
127 19
19.5 16.4
557 85
85.6 73.3
31.1 ± 8.7 31.7 ± 9.1
29.6 ± 10.6 31.7 ± 11.7
213 157
55.3 59.0
74 53
19.2 19.9
324 233
84.2 87.6
28.5 ± 5.431.6 ± 9.2-
29.0 ± 9.933.2 ± 13.3-
180 33
55.4 55.0
66 14
20.3 23.3
87 37
88.3 61.7
(%)
HIV(+) Patients
HIV(-) Patients
(n)
170
22.2
31.8 ± 8.4
29.8 ± 10.7
651 116
150 20
23.0 17.2
31.9 ± 8.7 31.1 ± 5.7
385 266
82 68
21.3 25.6
325 60
67 15
20.6 20.0
Patients (n)
(n)
767
• Comparing mean age of patients included and not included, p
= 0.09.
PULMONARY TUBERCULOSIS ANO HUMAN IMMUNODEFICIENCY VIRUS INFECTION
TABLE 3 MORTALITY DURING ANTITUBERCULOSIS THERAPY HIV(-)
HIV(+)
New cases Retreatments Total
(xln)
(%)
(xln)
(%)
371113"
32.7 27.0 31.3
16/444" 6/57 221501"
3.6 10.5 4.4
10/37 47/150"
Definition of abbreviations: x = number of patientswho died; n • number of IreaimenI charts reviewed. " p < 10" comparing HIVseroposilive wilh HIV seronegative patients.
tive patients (10/60,16.6070) than among HIV seronegative new cases (17/264, 6.4%; p = 0.02), and was, among HIV seropositive new cases, associated with a significantly higher therapy failure rate (p < 0.05). Therapy failures were not associated with the type of treatment received or with increasing age. Adverse events during treatment. The incidence of skin rashes was higher among HIV seropositive patients. Eighteen (27.3 %) of 66 HIV seropositive and 34 (12.1 %) of 281 HIV seronegative patients who started with streptomycin, isoniazid, and thiacetazone reported a rash during therapy (relative risk, 2.7; 95% CI, 1.4 to 5.5; p < 0.01). Thiacetazone was discontinued in 14 (77.8%) of 18 HIV seropositive and in 26 (76.5%) of 34 HIV seronegative patients who reported that skin rashes had occurred (p = 0.57). No cases of Stevens-Johnson syndrome were seen.
Follow-up after Completion of Chemotherapy Mortality after completion of chemotherapy. The mean duration of followup to assess mortality was similar for HIV seropositive (0.85 yr) and HIV
753
seronegative (0.90 yr) patients. The mortality of HIV seropositive patients (26.3/100patient-years) was significantly higher than that of HIV seronegative patients (2.21100 patient-years; p < 10-6 ; relative risk, 12.2; 95% CI, 3.9 to 40.3), and remained so in all subgroups when the patients were stratified as new caseslretreatments or by type of chemotherapy (table 5). Smear-positive pulmonary tuberculosis was still present immediately before to death in 2/15 HIV seropositive and in 215 HIV seronegative patients (p = 0.24).
Relapse rate ofpulmonary tuberculosis. The 42 patients diagnosed as therapy failure at the end of chemotherapy were excluded from this analysis. The average duration of the follow-up to assess the relapse rate of tuberculosis was shorter for the 58 HIV seropositive patients (0.75 yr; 95% CI, 0.67 to 0.84 yr) than for the 225 HIV seronegative patients (0.89 yr; 95% CI, 0.86 to 0.92 yr; p < 0.(01). This was mainly due to the increased mortality of the HIV seropositive patients. The relapse rate of pulmonary tuberculosis per 100 patient-years was 18.1 among HIV seropositive and 6.0 among HIV seronegative patients (relative risk, 3.0; 95% CI, 1.1 to 8.6; p < 0.04) (table 6). After stratification as new casesl retreatments at time of diagnosis, the difference in relapse rate between HIV seropositive and HIV seronegative patients was significant only among the more numerous new cases of tuberculosis. Relapse was not associated with the type of chemotherapy received, increasing age, or noncompliance with chemotherapy.
Overall tuberculosis treatment failure rate. The overall tuberculosis treatment failure rate among new cases of pulmonary tuberculosis was 21.1/100 patient-
TABLE 4 PULMONARY TUBERCULOSIS TREATMENT FAILURE RATE BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND TYPE OF ANTITUBERCULOSIS TREATMENT Type of PUlmonary Tuberculosis New cases Noncompliant Compliant Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
HIV (+) (xln) (%)
HIV(-) (x/n) (%)
1216020 6/1060 6/50 12 1/224.5
28/264 10.6 3/17 17.6 25/24710.1 1213930.8
0.07 0.06
33/231 14.3 6/51 11.8 1/21 4.5 40/303 13.2
NS NS NS NS
7/39 17.9 4/2714.8
2116 12.5 13/82 15.9
p Value
NS 0.014
Definition of abbreviations: x = number oflrealment failures; n = number evaluated;8TH and 8H: 8 = streptomycin, T = thiacetazone, I = isoniazid; RHEZ = rifampicin, isoniazid, ethambutol, and pyrazinamide.
years in HIV seropositive and 8.1/100 patient-years in HIV seronegative patients (relative risk, 2.6; 95% CI, 1.4 to 4.9; p = 0.(02). Discussion The present study confirmed that the mortality of HIV seropositive patients with pulmonary tuberculosis is very high, as it reached 31.3% during the year of chemotherapy and 26.3% during the year after chemotherapy. Among patients with pulmonary tuberculosis, "standard" antituberculosis therapy was less effective in HIV seropositive than in HIV seronegative patients because HIV seropositive patients tended to have a higher therapy failure rate and had a significantly increased relapse rate of pulmonary tuberculosis. Consequently, the overall tuberculosis therapy failure rate of new cases of pulmonary tuberculosis was 2.6 times higher among HIV seropositive patients than among HIV seronegative patients. In addition to immunodeficiency, the increased tuberculosis therapy failure rate in HIV seropositive patients could be ascribed in part to their lower compliance with chemotherapy. This was probably due to their poorer general health, related to other HIV-related symptoms, which made it more difficult for them to attend their treatment clinics, and not so much to concomitant therapy, as none received any antiretroviral treatment. The 14.3% tuberculosis therapy failure rate observed among HIV seronegative new cases of pulmonary tuberculosis treated with "standard" chemotherapy is also slightly higher than the 4 to 10% failure rate reported in earlier trials (22-24). A possible explanation is that HIV seronegative patients, cured at the end of therapy, may have been less likely to consult and therefore could not be included in the study. The induction of sputum production by nebulization of hypertonic saline (a technique unavailable in previous trials) and the high rates of primary resistance to antituberculosis drugs in Kinshasa among new cases of tuberculosis (32% resistant to isoniazid, 19% to streptomycin, 15% to thiacetazone (17]) may also have contributed to this observation. As relapses of pulmonary tuberculosis were diagnosed by sputum microscopyonly, this study undoubtedly underestimates their real incidence. Cases of less advanced paucibacillary pulmonary tuberculosis, reported to be more frequent in HIV seropositive patients, may
PERRI~NS, COLEBUNDERS, KARAHUNGA, ET AL.
754 TABLE 5
MORTALITY AFTER ANTITUBERCULOSIS THERAPY BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND ANTITUBERCULOSIS TREATMENT HIV (-)
HIV (+) Deaths! Patient-Years
Type of pulmonary tuberculosis New cases Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
Deaths! Patient-Years
(n)
(patients)
Deaths/100 Patient-Years
(n)
(patients)
Deaths!100 Patient-Years
8/38.9 7/18.3
48 19
20.6' 38.3t
4/202 1/30.2
225 33
1.98' 3.3t
7/27.5 4/16.7 4/12.9
32 20 15 67
25.5t 24.0* 31.0t 26.3§
21174.3
192 45 21 258
1.72t 4.9* ot 2.2§
15/57.1
2140.8 0/17.9 121232
For definition of abbreviations, see table 4. Comparing mortality of HIV seropositive and seronegative patients. • p < 10-'; relative risk, 10.4; 95% CI, 2.7 to 43.4. t p < 0.05. t p > 0.05. § P < 10"; relative risk, 5.1; 95% CI, 2.1 to 12.4.
TABLE 6 PULMONARY TUBERCULOSIS RELAPSE RATE BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND TYPE OF ANTITUBERCULOSIS TREATMENT HIV(-)
HIV (+) Relapsesl Patient-Years
Type of pulmonary tuberculosis New cases Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
Relapses!100 Patient-Years
(n)
(patients)
7/30.4 1113.7
40 18
23' 7.3
4/21.3 4/13.9 0/8.9 8/44.1
27 18 13 58
18.8 28.7 0 18.1t
Relapsesl Patient-Years (n)
(patients)
Relapses!100 Patient-Years
81180
202 23
4.4' 19.7
165 40 20 225
6.7 2.9 6.0 6.0t
4/20.3 10/148.5 1135 1/16.8
121200.3
For definition of abbreviations, see table 4. Comparing relapse rates between HIV seropositive and HIV seronegative patients . • p < 0.004; relative risk, 5.3; 95% CI, 1.6 to 17.5. t p = 0.04; relative risk, 3.0; 95% CI, 1.1 to 8.6.
have escaped detection (3, 13). Furthermore the study was not designed to diagnose extrapulmonary tu berculosis, even though extrapulmonary tuberculosis has been reported to be more common in HIV seropositive patients (1, 3-6, 9, 13). If these cases of less advanced pulmonary tuberculosis and extrapulmonary tuberculosis had been diagnosed, the discrepancy in results between HIV seropositive and HIV seronegative patients might even be greater than described. Skin rashes have been reported to occur more frequently in rifampicin-treated HIV seropositive patients (6). In this study HIV seropositive patients reported a 2.7-fold increased incidence of skin rashes over HIV seronegative patients when treated with "standard" antituberculosis chemotherapy. As thiacetazone was withheld in 77.80/0 of the patients who reported a skin rash, treatment per-
sonnel apparently associated these rash- therapy. However, the introduction of es with the use of thiacetazone. However, routine antituberculosis maintenance the 12% incidence of skin rashes report- therapy for HIV seropositive patients in ed by HIV seronegative patients treated African tuberculosis control programs is with thiacetazone in this study is higher likelyto be impossible because this would than the zero to 6% incidence reported necessitate screening all patients with from East and West African countries tuberculosis for HIV infection (26). As before the HIV epidemic (25). Therefore, short-course antituberculosis therapy is our data may be biased by inaccurate pa- more effective, and might result in a much tient recall or by an extremely prudent lower tuberculosis relapse rate, even in attitude of treatment personnel when HIV seropositive patients, and as its use confronted with minor cutaneous symp- has resulted in higherpatient compliance toms. Stevens-Johnsonsyndrome was not and case detection rates (27), future observed in this study, but it could easily studies should compare the public health have escaped detection, as detailed impact of its introduction with the confollow-up information during therapy tinued use of standard chemotherapy, folwas not available. lowed by maintenance therapy in HIV Given the high tuberculosis relapse rate seropositive patients. in HIV seropositive patients after "stanWeconclude that the mortality of HIV dard" chemotherapy, antituberculosis seropositive patients with pulmonary maintenance therapy in HIV seropositive tuberculosis treated with long-duration patients should be considered in HIV "standard" chemotherapy is significantseropositive patients who complete such ly higher than that of HIV seronegative
PULMONARY TUBERCULOSIS AND HUMAN IMMUNODEFICIENCY VIRUS INFECTION
patients with pulmonary tuberculosis receiving the same treatment. The efficacy of "standard" antituberculosis chemotherapy is lower in HIV seropositive than in HIV seronegative patients with pulmonary tuberculosis, mainly because HIV seropositive patients have a higher relapse rate of pulmonary tuberculosis than do HIV seronegative patients. Acknowledgment The writers thank the Zairian Ministry of Public Health for the authorization to publish the data. References 1. Pitchenik AE, Cole C, Russell BW, Fischl MA, Spira TJ, Snider DE. Tuberculosis, atypical mycobacteriosis, and the acquired immunodeficiency syndrome among Haitian and non-Haitian patients in South Florida. Ann Intern Med 1984; 101:641-5. 2. Centers for Disease Control. Tuberculosis and the acquired immunodeficiency syndrome, Florida. MMWR 1986; 35:587-90. 3. Sunderham G, McDonald RJ, Maniatis T, Oleske J, Kapila R, Reichman LB. Tuberculosis as a manifestation of the acquired immunodeficiency syndrome. JAMA 1986; 256:362-6. 4. Pitchenik AE, Burr J, Suarez M, Fertel D, Gonzalez G, Moas C. Human T-celllymphocytotrophic virus-III seropositivity and related disease among 71 consecutive patients in whom tuberculosis was diagnosed. Am Rev Respir Dis 1987; 135:875-9. 5. Chaisson RE, Schecter GA, Theuer CP, Rutherford GW, Echenberg DF, Hopewell PC. Tuberculosis in patients with the acquired immunodeficiency syndrome: clinical features, response to therapy, and survival. Am Rev Respir Dis 1987; 136:570-4. 6. Sonnet J, Prignot J, Zech F, Willame JC. High prevalence of tuberculosis in HIV infected black
755
patients from Central Africa. Ann Soc BeIge Med Trop 1987; 67:299-300. 7. Centers for Disease Control. Tuberculosis, final data, United States, 1986. MMWR 1986; 36:817-9. 8. Mann J, Snider DE, Francis H, Quinn T, Colebunders R, Piot P, Curran J, Nzila N, Ngaly B, Matiatudila M, Dikilu K, Masaki Mu Nzingg, Nkoko B. Association between HTLV-III/LAV Infection and tuberculosis in Zaire. JAMA 1986; 256:346. 9. Colebunders RL, Ryder R, Nzila N, Nkoko B, Jeugmans J, Kalala M, Francis H, Mann J, Quinn T, Piot P. HIV infection in patients with tuberculosis in Kinshasa, Zaire. Am Rev Respir Dis 1989; 139:1082-5. 10. Standaert B, Niragira F, Kadende P, Piot P. The association of tuberculosis and HIVinfection in Burundi. AIDS Res Hum Retroviruses 1989; 5:247-51. 11. Meeran K. Prevalence of HIV infection among patients with leprosy and tuberculosis in rural zambia. Br Med J 1989; 298:364-5. 12. Willame JC, Nkoko B, Pauwels P, et at. Tuberculose et seropositivite anti-VIH Kinshasa, Zaire. Ann Soc BeIge Med Trop 1988; 68:165-7. 13. Selwyn P, Hartel D, Lewis V, et at. A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N Engl J Med 1989; 320:545-50. 14. Slutkin G, Leowski J, Mann J. The effects of the AIDS epidemic on the tuberculosis problem and tuberculosis programmes. Bull Int Union Tuberc Lung Dis 1988; 63:21-4. 15. Willame JC, Prignot J, Masaki Ma Nsinga, Kaboto M. Enquete tuberculinique et radiophotographique transversale sur l'endemie tuberculeuse Kinshasa, Zaire (1978-1979). Ann Soc BeIge Med Trop 1981; 61:489-501. 16. N'Galy B, Ryder RW. Epidemiology of HIV infection in Africa. J Acquir Immune Defic Syndr 1988; 1:551-8. 17. Bureau National de la Tuberculose, Programme National anti-tuberculeux inregre aux soins
a
a
de sante de base, 1989. Ed. Departement de la Sante Publique, Kinshasa, Zaire. 18. Anonymous. Les mycobacteries, Ed. Institut Pasteur Production. Marnes la Coquette, 1981. 19. Pattyn S, Portaels F. In vitro cultivation and characterisation of Mycobacterium /epraemurium. Int J Lepr 1980; 48:7-14. 20. Jenkins P, Pattyn S, Portaels F. Diagnostic bacteriology. In: Ratledge C, Stanford JC, eds. The biology of the mycobacteria. London: Academic Press; 1982;441-71. 21. Canetti G, Froman S, Grosset J, et at. Mycobacteria: laboratory methods for testing drug sensitivity and resistance. Bull WHO 1963; 29:565-78. 22. East African/British Medical Research Council Third Thiacetazone Investigation. Isoniazid with thiacetazone (thioacetazone) in the treatment of pulmonary tuberculosis in East Africa. Third investigation: The effect of an initial streptomycin supplement. Tubercle 1966; 47:1-32. 23. East African/British Medical Reseach Council Fifth Thiacetazone Investigation. Isoniazid with thiacetazone (thioacetazone) in the treatement of pulmonary tuberculosis in East Africa. Fifth Investigation. Tubercle 1970; 51:123-51. 24. Tuberculosis Chemotherapy Centre, Madras. A concurrent comparison of intermittent (twice weekly) isoniazid plus streptomycin and daily isoniazid plus PAS in the domiciliary treatment of pulmonary tuberculosis. Bull WHO 1964; 31:247-71. 25. Miller A, Fox W, ThII R. An international cooperative investigation into thiacetazone (thioacetazone) side-effects. Tubercle 1966; 47:33-73. 26. Colebunders R, Braun M, Nzila N, Dikilu K, Muepu K, Ryder R. Evaluation of the World Health Organisation clinical case definition of AIDS among tuberculosis patients in Kinshasa, Zaire. J Infect Dis 1989; 160:902-3. 27. Styblo K. Overview and epidemiological assessment of the current global tuberculosis situation: with an emphasis on tuberculosis control in developing countries. Bull Int Union Tuberc Lung Dis 1988; 63:39-44.
JOSEPH H. PERRIENS, ROBERT L. COLEBUNDERS, CIBANGU KARAHUNGA, JEAN-CLAUDE WILLAME, JACQUES JEUGMANS, MULAMBA KABOTO, YIADIUL MUKADI, PAUL PAUWELS, ROBERT W. RYDER, JACQUES PRIGNOT, and PETER PlOT
Introduction
Many reports from both industrialized and developing countries have demonstrated a strong association between tuberculosis and human immunodeficiency virus (HIV) infection (1-12). Patients with inactive tuberculosis infection who subsequently acquire HIV infection have been shown to be at greatly increased risk of developing active tuberculosis (13). Consequently, there is a general consensus that the HIV epidemic will greatly complicate tuberculosis control (14). Antituberculosis chemotherapy is one of the mainstays of tuberculosis control, and chemotherapy including rifampin is able to achieve similar antimicrobial efficacy in patients with and without HIV infection (3, 5). There are, however,conflicting data on the efficacy of "standard" long-duration antituberculosis therapy regimens, including streptomycin, isoniazid, and thiacetazone, which are still in widespread use in many developingcountries (9, 10). The aim of the present study was to assess the treatment outcomes of HIV seropositive and HIV seronegative patients with pulmonary tuberculosis treated according to a streptomycin-based treatment policy. Methods The study wasconducted in Kinshasa, the capital city of zaire, where 70070 of the adult population is purified protein derivative positive, where the incidence of pulmonary tuberculosis is estimated to be 3/1,000 inhabitants/year, and where 3 to 8% of the general adult population is estimated to be HIV seropositive (15, 16). 750
SUMMARY Toevaluate their treatment outcomes 170human Immunodeficiency virus (HIV) seropositive and 597 HIV seronegative patients with active pulmonary tuberculosis (TB) treated for 1 yr with "stsndard" chemotherapy, Including streptomycin, Isoniazid, and, In most cases, thiacetazone, were traced at completion of therapy. All 582 survivors were Invited for reevaluation, and 385 patients, of whom 82 (21.3%) were HIV seropositive, were evaluated. Of those, 325 consenting patients, of whom fJ1 (20.6%) were HIV seropositive, were followed for 12 months. One year after TB had been diagnosed 47 (31.3%) of the 150 HIV seroposltlva and 22 (4.4%) of the 501 HIV seronegatlva patients traced had died (p 10-'). During the subsequent year the mortsllty of fJ1 HIV seropositive patients (26.31100 patlent-years) was higher than that of the 303 HIV seronegative patients (2.21100 patlents-years, p = 10-'). HIV seropositive patients hed a higher overall TB therapy failure rate 24 months after the diagnosis of TB than did HIV seronegative patients (21.11100 patlent-years versus 8.11100 patient-years, p 0.002), mainly because their relapse rate of pulmonary TB (18.11100 patient-years) wes higher than that of HIV seronegative patients (6.01100 patient-years, p 0.03). Given their higher relapse rate after 1 yr of "stendard" chemotherapy, the public health Impact of routine maintenance therapy In HIV seropositive patients with pulmonary TB who complete such therapy should be assessed In comparison to the Introduction of rlfamplcln-based short-eourse antituberculosis chemotherapy In developing countries. AM REV RESPIR DIS 1991; 144:750-755
=
=
Whenever feasible patients with tuberculosis are treated as outpatients in primary health care centers. "Standard" chemotherapy for pulmonary tuberculosis in adults weighing more than 50 kg consists of daily intramuscularly administered streptomycin (l g) and orally administered thiacetazone (150 mg) plus isoniazid (300 mg) for 2 months, followed by daily orally administered thiacetazone (150 mg) plus isoniazid (300 mg) for 10 months. If thiacetazone is unavailable, if the patient refuses to use it, or if side effects to thiacetazone occur, daily streptomycin (1 g) and isoniazid (300 mg) for 2 months, followed by isoniazid (900 mg) twice a week plus streptomycin (1 g) for 10 months is given instead (17). Therapy failures and relapses of pulmonary tuberculosis occurring within 12 months of completion of chemotherapy receivea retreatment scheme that consists of 2 months daily rifampicin (600 mg), ethambutol (1,200 mg), isoniazid (600 mg), and pyrazinamide (2,000
=
mg), followed by 4 months rifampicin (600 mg) and isoniazid (900 mg) twice a week (17). All study patients were treated according to this treatment policy. None of the HIV
(Receivedin originalform August 13, 1990and in revised form April 16, 1991) 1 From Projet Sida, the Bureau National de la Thberculose, the Centre de Depistage de la Thberculose, Kinshasa, Zaire; the Institute of Tropical Medicine, Antwerp, the Belgo-Zairian Medical Cooperation, Brussels, and the Universite Catholique de Louvain, Mont Godinne, Belgium; the Center for Infectious Disease, Centers for Disease Control, Atlanta, Georgia. 2 Supported by the Belgo-Zairian Medical Cooperation and by the Banque Commerciale du zaire. 3 Correspondence and requests for reprints should be addressed to J. Perriens, Institute of 'Iropical Medicine, Department of Microbiology, Nationalestraat 155, B-2000 Antwerp, Belgium.
PULMONARY TUBERCULOSIS AND HUMAN IMMUNODEFICIENCY VIRUS INFECTION
seropositive patients was treated with antiretroviral drugs.
Study Design The study was conducted in four phases. (1) Patient identification: HIVserosurveys among patients with tuberculosis. In 1987, t '.voHIV serosurveys were performed in Kin-
shasa among inpatients and outpatients with tuberculosis (9, 12). The HIV serosurvey among inpatients was done in May 1987 in a sanatorium. Outpatients with smear-positive pulmonary tuberculosis were studied in Kinshasa's biggest tuberculosis diagnosis center, the Centre de Depistage de la Tuberculose, between January and April 1987.
(2) Chart review at the end ofantituberculosis chemotherapy. One year later, between January 1988 and May 1988, when the patients identified by the HIV serosurveys had completed antituberculosis therapy, their treatment charts were reviewed, and mortality during treatment was recorded.
(3) Patient evaluation at the end of antituberculosis therapy. Surviving patients, whose treatment charts were found, and who could be physically located, were invited for reevaluation (between January and May 1988). Patients who showed up for evaluation were interviewed concerning their symptoms, and a physical examination, repeat HIV serology, sputum microscopy, and sputum culture were performed. (4) Follow-up after completion of antituberculosis therapy. Consenting patients,
evaluated at the end of chemotherapy, were enrolled in a longitudinal study, which lasted from January 1988to May 1989. In this study, history, physical examination, and sputum microscopy were repeated bimonthly. Resources to do sputum cultures were not available in this phase of the study. Mortality was recorded.
Diagnostic Procedures Sputum microscopy was done on freshly expectorated sputum, induced by nebulization of hypertonic 3070 saline if sputum could not be spontaneously produced, homogenized, and decontaminated using the sodium laurylsulfate method (18), and centrifuged for 30 min at 3,000 x g. Ziehl-Neelsen-stained smears of the sediment were examined under light microscopy by one of us (P. Pauwels). Sputum cultures were performed in duplicate on Lcwenstein-Jensen medium (Bacto Loewenstein Medium Base; Difco Laboratories, Detroit, MI) and Ogawa medium (19). The mycobacterial strains were identified using nitrate production, niacin production, and thiophene carboxylic acid hydrazide susceptibility tests as described by Jenkins and coworkers (20). Susceptibility tests for streptomycin, isoniazid, thiacetazone, ethambutol, and rifampicin were done on LOwenstein media using the absolute concentration method (21). HIV-serology was done in duplicate using an ELISA test (Wellcozyme; Wellcome Diagnostics, Dart ford, UK, or Organon Vironostica, Organon Teknika, Oss, The
Netherlands). Repeatedly reactive samples by ELISA were confirmed by indirect immunofluorescence in the HIV serosurvey among outpatients, and by Western blot (Dupont de Nemours, Wilmington, DE) in the serosurvey among inpatients and at the end of chemotherapy. A Western blot was considered positive if antibodies against p24 and gp41 and/or gp120/160 were present.
Definitions Patients with sputum smears positive for acidfast bacilli or sputum cultures positive for Mycobacterium tuberculosis were considered to have active pulmonary tuberculosis. Smear-positive patients and patients with sputum cultures positive for M tuberculosis 12 months after starting chemotherapy were considered therapy failures. Patients who at reevaluation after chemotherapy had sputum smears and sputum cultures negative for mycobacteria, and who later had sputum smears positive for acid-fast bacilli, were considered to have a relapse of pulmonary tuberculosis. Patients who initially (at the time the HIV serosurveys were done) had started treatment for a therapy failure or a relapse will be called "retreatments" in this report. Other patients will be called "new cases." The overall tuberculosis therapy failure rate was considered to be the therapy failure rate plus the relapse rate. A patient was considered compliant with therapy if he had completed at least 11 of the 12 months therapy prescribed, at least 56 of the 60 streptomycin injections prescribed during the first 2 months of chemotherapy, and if he had never interrupted therapy for more than 2 consecutive weeks.
Statistical Analysis The chi-square test (with Yates correction in 2 x 2 tables), Fischer's exact test, the t test, and the Mann-Whitney test were used as appropriate. The significance level used was p < 0.05.
Results
Patient Population The patient flow through the study is outlined in table 1,and characteristics of patients in/excluded in subsequent phases of the study are outlined in table 2. A total of 767 patients with microbiologically confirmed pulmonary tuberculosis was examined in the HIV serosurveys, of whom 170 (22.20/0) were HIV seropositive. HIV seroprevalence was lower among outpatients (124/621, 20%) than among inpatients (46/146, 31.5%; p < 0.01), and lower among new cases (124/642, 19.3%) than among retreatments (46/125,36.8%; p < 0.01). Female patients were more often HIV seropositive (87/319,27.3%) than were male patients (83/448, 18.5%; p = 0.005). HIV
751
seropositive patients were on average somewhat older (mean, 31.8 yr; SO, 8.4 yr) than were HIV seronegative patients (mean, 29.8 yr; SO, 10.7 yr; p = 0.001). Demographic characteristics of inpatients and outpatients were similar. The characteristics of patients included in the chart review after therapy, of patients who were subsequently reevaluated, and of patients who werethereafter followed, were similar to the ones of patients included in the HIV serosurveys for age and sex, the distribution of newcases and retreatments, and the distribution of inpatients and outpatients. Of the patients known to be alive after the chart review was completed, a higher proportion of HIV seropositive (82/103, 79.6%) than of HIV seronegative patients (303/479,63.3%) was subsequently reevaluated (p < 0.002). This was a consequence of more intensive follow-up efforts aimed at the HIV seropositive patients. Consequently, the proportion of HIV seropositive patients remained constant throughout the study despite the higher mortality of HIV seropositive patients during therapy. Nine patients, seven seroconversion and two initially HIV seropositive who were HIV seronegative on reevaluation, were excluded from the study, as the original samples from HIV serosurveys were not available for retesting.
Chart Review: Mortality during the Year after the Start of Antituberculosis Chemotherapy The one-year mortality among HIV seropositivepatients was 31.30/0 (47/150), significantly higher than the 22/501 (4.4%) among HIV seronegativepatients (relative risk, 7.0; 95% CI, 4.0 to 12.6; p < 10-6 ) (table 3). The mortality of HIV seropositive new cases and retreatments from both serosurveys was similar, but among HIV seronegative patients, retreatments had a higher mortality rate (10.5%) than among new cases (3.6%; p < 0.05). Patient Evaluation at the End of Chemotherapy Efficacy ofantituberculosis therapy. At the end of chemotherapy there were 13 (15.9%) therapy failures among 82 HIV seropositive patients and 40 (13.2%) among 303 HIV seronegative patients (p = 0.65) (table 4). Of the 53 therapy failures, diagnosis in two (3.7%) was made by sputum smear only, eight (15.1%) were smear positive and culture positive.and 43 (81.1 %) wereculture positive only.
PERRI~NS, COLEBUNDERS, KARAHUNGA, ET AL.
752 TABLE 1
PATIENT FLOW BY HUMAN IMMUNODEFICIENCY VIRUS (HIV) SEROLOGIC STATUS HIV (-)
HIV (+) Patient identification: HIV serosurveys (January-May 1987)
46 Inpatients 124 Outpatients
100 Inpatients 497 Outpatients
170 Total
597 Total
Chart not found: 20 Chart found: 150
Chart review at the end of chemotherapy (January-May 1988)
Chart not found: 96 Chart found: 501
IDead: 22
IDead: 47 Alive: 103
Alive: 479 Excluded: 7
Excluded: 2
Patient evaluation at end of chemotherapy (January-May 1988)
Not evaluated: 19 IEvaluated: 82
INot evaluated:
ITB not cured:
ITB not cured: 40 TB cured: 263
13
TB cured: 69 Longitudinal study after end of therapy (January 1988-May 1989)
Follow up on relapse
Refused FU: 11 Included in FU: 58 Relapsed: 8
No relapse: 50 (including 26 with FU < 12 months)
169
Evaluated: 303
Follow up on mortality
Refused FU: 15 Included in FU: 67
Follow up on relapse
Follow up on mortality
Refused FU: 38 Included in FU: 225
Refused FU: 45 Included in FU: 258
Died: 15
Died: 5
Relapsed: 12
Alive: 52 (including 29 with FU < 12 months)
Alive: 253 (including 53 with FU < 12 months)
No relapse: 213 (inclUding 49 with FU < 12 months)
Definition of abbreviations: FU = follow up; TB = tu berculosis.
The only mycobacterium isolated from the 51patients with positive sputum cultures at the end of chemotherapy was M tuberculosis. Susceptibility tests were done on 50 isolates. Only three were sensitiveto all fivedrugs tested. Isolates from HIV seropositive patients were resistant
to fewer drugs (on average, 1.5 drugs; 950/0 CI, 0.9 to 2.1) than isolates from HIV seronegative patients (on average, 2.5 drugs; 95% CI, 2.2 to 2.8; p = 0.003). Resistance to individual antimicrobial agents as 48% for streptomycin, 92% for isoniazid, 68% for thiacetazone, 6% for
rifampicin, and 12% for ethambutol. As susceptibility tests were not performed at the start of therapy, the causal relationship of resistance with the occurrence of therapy failures could not be analyzed. Noncompliance with chemotherapy was more frequent among HIV seroposi-
TABLE 2 CHARACTERISTICS OF THE PATIENT POPULATION
Patient identification Chart review at the end of chemotherapy Included Lost Reevaluation at the end of chemotherapy Included Not included Follow-up on mortality after chemotherapy Included Not included
HIV(+) Patients
Mean ± SO Age at Diagnosis
Male Patients
Ex-inpatients
New Cases
(%)
(n)
(%)
(n)
(%)
448
58.4
146
19.0
642
83.6
29.6 ± 10.7 30.7 ± 10.7
370 78
56.8 67.2
127 19
19.5 16.4
557 85
85.6 73.3
31.1 ± 8.7 31.7 ± 9.1
29.6 ± 10.6 31.7 ± 11.7
213 157
55.3 59.0
74 53
19.2 19.9
324 233
84.2 87.6
28.5 ± 5.431.6 ± 9.2-
29.0 ± 9.933.2 ± 13.3-
180 33
55.4 55.0
66 14
20.3 23.3
87 37
88.3 61.7
(%)
HIV(+) Patients
HIV(-) Patients
(n)
170
22.2
31.8 ± 8.4
29.8 ± 10.7
651 116
150 20
23.0 17.2
31.9 ± 8.7 31.1 ± 5.7
385 266
82 68
21.3 25.6
325 60
67 15
20.6 20.0
Patients (n)
(n)
767
• Comparing mean age of patients included and not included, p
= 0.09.
PULMONARY TUBERCULOSIS ANO HUMAN IMMUNODEFICIENCY VIRUS INFECTION
TABLE 3 MORTALITY DURING ANTITUBERCULOSIS THERAPY HIV(-)
HIV(+)
New cases Retreatments Total
(xln)
(%)
(xln)
(%)
371113"
32.7 27.0 31.3
16/444" 6/57 221501"
3.6 10.5 4.4
10/37 47/150"
Definition of abbreviations: x = number of patientswho died; n • number of IreaimenI charts reviewed. " p < 10" comparing HIVseroposilive wilh HIV seronegative patients.
tive patients (10/60,16.6070) than among HIV seronegative new cases (17/264, 6.4%; p = 0.02), and was, among HIV seropositive new cases, associated with a significantly higher therapy failure rate (p < 0.05). Therapy failures were not associated with the type of treatment received or with increasing age. Adverse events during treatment. The incidence of skin rashes was higher among HIV seropositive patients. Eighteen (27.3 %) of 66 HIV seropositive and 34 (12.1 %) of 281 HIV seronegative patients who started with streptomycin, isoniazid, and thiacetazone reported a rash during therapy (relative risk, 2.7; 95% CI, 1.4 to 5.5; p < 0.01). Thiacetazone was discontinued in 14 (77.8%) of 18 HIV seropositive and in 26 (76.5%) of 34 HIV seronegative patients who reported that skin rashes had occurred (p = 0.57). No cases of Stevens-Johnson syndrome were seen.
Follow-up after Completion of Chemotherapy Mortality after completion of chemotherapy. The mean duration of followup to assess mortality was similar for HIV seropositive (0.85 yr) and HIV
753
seronegative (0.90 yr) patients. The mortality of HIV seropositive patients (26.3/100patient-years) was significantly higher than that of HIV seronegative patients (2.21100 patient-years; p < 10-6 ; relative risk, 12.2; 95% CI, 3.9 to 40.3), and remained so in all subgroups when the patients were stratified as new caseslretreatments or by type of chemotherapy (table 5). Smear-positive pulmonary tuberculosis was still present immediately before to death in 2/15 HIV seropositive and in 215 HIV seronegative patients (p = 0.24).
Relapse rate ofpulmonary tuberculosis. The 42 patients diagnosed as therapy failure at the end of chemotherapy were excluded from this analysis. The average duration of the follow-up to assess the relapse rate of tuberculosis was shorter for the 58 HIV seropositive patients (0.75 yr; 95% CI, 0.67 to 0.84 yr) than for the 225 HIV seronegative patients (0.89 yr; 95% CI, 0.86 to 0.92 yr; p < 0.(01). This was mainly due to the increased mortality of the HIV seropositive patients. The relapse rate of pulmonary tuberculosis per 100 patient-years was 18.1 among HIV seropositive and 6.0 among HIV seronegative patients (relative risk, 3.0; 95% CI, 1.1 to 8.6; p < 0.04) (table 6). After stratification as new casesl retreatments at time of diagnosis, the difference in relapse rate between HIV seropositive and HIV seronegative patients was significant only among the more numerous new cases of tuberculosis. Relapse was not associated with the type of chemotherapy received, increasing age, or noncompliance with chemotherapy.
Overall tuberculosis treatment failure rate. The overall tuberculosis treatment failure rate among new cases of pulmonary tuberculosis was 21.1/100 patient-
TABLE 4 PULMONARY TUBERCULOSIS TREATMENT FAILURE RATE BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND TYPE OF ANTITUBERCULOSIS TREATMENT Type of PUlmonary Tuberculosis New cases Noncompliant Compliant Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
HIV (+) (xln) (%)
HIV(-) (x/n) (%)
1216020 6/1060 6/50 12 1/224.5
28/264 10.6 3/17 17.6 25/24710.1 1213930.8
0.07 0.06
33/231 14.3 6/51 11.8 1/21 4.5 40/303 13.2
NS NS NS NS
7/39 17.9 4/2714.8
2116 12.5 13/82 15.9
p Value
NS 0.014
Definition of abbreviations: x = number oflrealment failures; n = number evaluated;8TH and 8H: 8 = streptomycin, T = thiacetazone, I = isoniazid; RHEZ = rifampicin, isoniazid, ethambutol, and pyrazinamide.
years in HIV seropositive and 8.1/100 patient-years in HIV seronegative patients (relative risk, 2.6; 95% CI, 1.4 to 4.9; p = 0.(02). Discussion The present study confirmed that the mortality of HIV seropositive patients with pulmonary tuberculosis is very high, as it reached 31.3% during the year of chemotherapy and 26.3% during the year after chemotherapy. Among patients with pulmonary tuberculosis, "standard" antituberculosis therapy was less effective in HIV seropositive than in HIV seronegative patients because HIV seropositive patients tended to have a higher therapy failure rate and had a significantly increased relapse rate of pulmonary tuberculosis. Consequently, the overall tuberculosis therapy failure rate of new cases of pulmonary tuberculosis was 2.6 times higher among HIV seropositive patients than among HIV seronegative patients. In addition to immunodeficiency, the increased tuberculosis therapy failure rate in HIV seropositive patients could be ascribed in part to their lower compliance with chemotherapy. This was probably due to their poorer general health, related to other HIV-related symptoms, which made it more difficult for them to attend their treatment clinics, and not so much to concomitant therapy, as none received any antiretroviral treatment. The 14.3% tuberculosis therapy failure rate observed among HIV seronegative new cases of pulmonary tuberculosis treated with "standard" chemotherapy is also slightly higher than the 4 to 10% failure rate reported in earlier trials (22-24). A possible explanation is that HIV seronegative patients, cured at the end of therapy, may have been less likely to consult and therefore could not be included in the study. The induction of sputum production by nebulization of hypertonic saline (a technique unavailable in previous trials) and the high rates of primary resistance to antituberculosis drugs in Kinshasa among new cases of tuberculosis (32% resistant to isoniazid, 19% to streptomycin, 15% to thiacetazone (17]) may also have contributed to this observation. As relapses of pulmonary tuberculosis were diagnosed by sputum microscopyonly, this study undoubtedly underestimates their real incidence. Cases of less advanced paucibacillary pulmonary tuberculosis, reported to be more frequent in HIV seropositive patients, may
PERRI~NS, COLEBUNDERS, KARAHUNGA, ET AL.
754 TABLE 5
MORTALITY AFTER ANTITUBERCULOSIS THERAPY BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND ANTITUBERCULOSIS TREATMENT HIV (-)
HIV (+) Deaths! Patient-Years
Type of pulmonary tuberculosis New cases Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
Deaths! Patient-Years
(n)
(patients)
Deaths/100 Patient-Years
(n)
(patients)
Deaths!100 Patient-Years
8/38.9 7/18.3
48 19
20.6' 38.3t
4/202 1/30.2
225 33
1.98' 3.3t
7/27.5 4/16.7 4/12.9
32 20 15 67
25.5t 24.0* 31.0t 26.3§
21174.3
192 45 21 258
1.72t 4.9* ot 2.2§
15/57.1
2140.8 0/17.9 121232
For definition of abbreviations, see table 4. Comparing mortality of HIV seropositive and seronegative patients. • p < 10-'; relative risk, 10.4; 95% CI, 2.7 to 43.4. t p < 0.05. t p > 0.05. § P < 10"; relative risk, 5.1; 95% CI, 2.1 to 12.4.
TABLE 6 PULMONARY TUBERCULOSIS RELAPSE RATE BY HIV SEROSTATUS, TYPE OF TUBERCULOSIS AT DIAGNOSIS, AND TYPE OF ANTITUBERCULOSIS TREATMENT HIV(-)
HIV (+) Relapsesl Patient-Years
Type of pulmonary tuberculosis New cases Retreatments Type of antituberculosis treatment STH STH followed by SH or RHEZ SH or RHEZ Total
Relapses!100 Patient-Years
(n)
(patients)
7/30.4 1113.7
40 18
23' 7.3
4/21.3 4/13.9 0/8.9 8/44.1
27 18 13 58
18.8 28.7 0 18.1t
Relapsesl Patient-Years (n)
(patients)
Relapses!100 Patient-Years
81180
202 23
4.4' 19.7
165 40 20 225
6.7 2.9 6.0 6.0t
4/20.3 10/148.5 1135 1/16.8
121200.3
For definition of abbreviations, see table 4. Comparing relapse rates between HIV seropositive and HIV seronegative patients . • p < 0.004; relative risk, 5.3; 95% CI, 1.6 to 17.5. t p = 0.04; relative risk, 3.0; 95% CI, 1.1 to 8.6.
have escaped detection (3, 13). Furthermore the study was not designed to diagnose extrapulmonary tu berculosis, even though extrapulmonary tuberculosis has been reported to be more common in HIV seropositive patients (1, 3-6, 9, 13). If these cases of less advanced pulmonary tuberculosis and extrapulmonary tuberculosis had been diagnosed, the discrepancy in results between HIV seropositive and HIV seronegative patients might even be greater than described. Skin rashes have been reported to occur more frequently in rifampicin-treated HIV seropositive patients (6). In this study HIV seropositive patients reported a 2.7-fold increased incidence of skin rashes over HIV seronegative patients when treated with "standard" antituberculosis chemotherapy. As thiacetazone was withheld in 77.80/0 of the patients who reported a skin rash, treatment per-
sonnel apparently associated these rash- therapy. However, the introduction of es with the use of thiacetazone. However, routine antituberculosis maintenance the 12% incidence of skin rashes report- therapy for HIV seropositive patients in ed by HIV seronegative patients treated African tuberculosis control programs is with thiacetazone in this study is higher likelyto be impossible because this would than the zero to 6% incidence reported necessitate screening all patients with from East and West African countries tuberculosis for HIV infection (26). As before the HIV epidemic (25). Therefore, short-course antituberculosis therapy is our data may be biased by inaccurate pa- more effective, and might result in a much tient recall or by an extremely prudent lower tuberculosis relapse rate, even in attitude of treatment personnel when HIV seropositive patients, and as its use confronted with minor cutaneous symp- has resulted in higherpatient compliance toms. Stevens-Johnsonsyndrome was not and case detection rates (27), future observed in this study, but it could easily studies should compare the public health have escaped detection, as detailed impact of its introduction with the confollow-up information during therapy tinued use of standard chemotherapy, folwas not available. lowed by maintenance therapy in HIV Given the high tuberculosis relapse rate seropositive patients. in HIV seropositive patients after "stanWeconclude that the mortality of HIV dard" chemotherapy, antituberculosis seropositive patients with pulmonary maintenance therapy in HIV seropositive tuberculosis treated with long-duration patients should be considered in HIV "standard" chemotherapy is significantseropositive patients who complete such ly higher than that of HIV seronegative
PULMONARY TUBERCULOSIS AND HUMAN IMMUNODEFICIENCY VIRUS INFECTION
patients with pulmonary tuberculosis receiving the same treatment. The efficacy of "standard" antituberculosis chemotherapy is lower in HIV seropositive than in HIV seronegative patients with pulmonary tuberculosis, mainly because HIV seropositive patients have a higher relapse rate of pulmonary tuberculosis than do HIV seronegative patients. Acknowledgment The writers thank the Zairian Ministry of Public Health for the authorization to publish the data. References 1. Pitchenik AE, Cole C, Russell BW, Fischl MA, Spira TJ, Snider DE. Tuberculosis, atypical mycobacteriosis, and the acquired immunodeficiency syndrome among Haitian and non-Haitian patients in South Florida. Ann Intern Med 1984; 101:641-5. 2. Centers for Disease Control. Tuberculosis and the acquired immunodeficiency syndrome, Florida. MMWR 1986; 35:587-90. 3. Sunderham G, McDonald RJ, Maniatis T, Oleske J, Kapila R, Reichman LB. Tuberculosis as a manifestation of the acquired immunodeficiency syndrome. JAMA 1986; 256:362-6. 4. Pitchenik AE, Burr J, Suarez M, Fertel D, Gonzalez G, Moas C. Human T-celllymphocytotrophic virus-III seropositivity and related disease among 71 consecutive patients in whom tuberculosis was diagnosed. Am Rev Respir Dis 1987; 135:875-9. 5. Chaisson RE, Schecter GA, Theuer CP, Rutherford GW, Echenberg DF, Hopewell PC. Tuberculosis in patients with the acquired immunodeficiency syndrome: clinical features, response to therapy, and survival. Am Rev Respir Dis 1987; 136:570-4. 6. Sonnet J, Prignot J, Zech F, Willame JC. High prevalence of tuberculosis in HIV infected black
755
patients from Central Africa. Ann Soc BeIge Med Trop 1987; 67:299-300. 7. Centers for Disease Control. Tuberculosis, final data, United States, 1986. MMWR 1986; 36:817-9. 8. Mann J, Snider DE, Francis H, Quinn T, Colebunders R, Piot P, Curran J, Nzila N, Ngaly B, Matiatudila M, Dikilu K, Masaki Mu Nzingg, Nkoko B. Association between HTLV-III/LAV Infection and tuberculosis in Zaire. JAMA 1986; 256:346. 9. Colebunders RL, Ryder R, Nzila N, Nkoko B, Jeugmans J, Kalala M, Francis H, Mann J, Quinn T, Piot P. HIV infection in patients with tuberculosis in Kinshasa, Zaire. Am Rev Respir Dis 1989; 139:1082-5. 10. Standaert B, Niragira F, Kadende P, Piot P. The association of tuberculosis and HIVinfection in Burundi. AIDS Res Hum Retroviruses 1989; 5:247-51. 11. Meeran K. Prevalence of HIV infection among patients with leprosy and tuberculosis in rural zambia. Br Med J 1989; 298:364-5. 12. Willame JC, Nkoko B, Pauwels P, et at. Tuberculose et seropositivite anti-VIH Kinshasa, Zaire. Ann Soc BeIge Med Trop 1988; 68:165-7. 13. Selwyn P, Hartel D, Lewis V, et at. A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N Engl J Med 1989; 320:545-50. 14. Slutkin G, Leowski J, Mann J. The effects of the AIDS epidemic on the tuberculosis problem and tuberculosis programmes. Bull Int Union Tuberc Lung Dis 1988; 63:21-4. 15. Willame JC, Prignot J, Masaki Ma Nsinga, Kaboto M. Enquete tuberculinique et radiophotographique transversale sur l'endemie tuberculeuse Kinshasa, Zaire (1978-1979). Ann Soc BeIge Med Trop 1981; 61:489-501. 16. N'Galy B, Ryder RW. Epidemiology of HIV infection in Africa. J Acquir Immune Defic Syndr 1988; 1:551-8. 17. Bureau National de la Tuberculose, Programme National anti-tuberculeux inregre aux soins
a
a
de sante de base, 1989. Ed. Departement de la Sante Publique, Kinshasa, Zaire. 18. Anonymous. Les mycobacteries, Ed. Institut Pasteur Production. Marnes la Coquette, 1981. 19. Pattyn S, Portaels F. In vitro cultivation and characterisation of Mycobacterium /epraemurium. Int J Lepr 1980; 48:7-14. 20. Jenkins P, Pattyn S, Portaels F. Diagnostic bacteriology. In: Ratledge C, Stanford JC, eds. The biology of the mycobacteria. London: Academic Press; 1982;441-71. 21. Canetti G, Froman S, Grosset J, et at. Mycobacteria: laboratory methods for testing drug sensitivity and resistance. Bull WHO 1963; 29:565-78. 22. East African/British Medical Research Council Third Thiacetazone Investigation. Isoniazid with thiacetazone (thioacetazone) in the treatment of pulmonary tuberculosis in East Africa. Third investigation: The effect of an initial streptomycin supplement. Tubercle 1966; 47:1-32. 23. East African/British Medical Reseach Council Fifth Thiacetazone Investigation. Isoniazid with thiacetazone (thioacetazone) in the treatement of pulmonary tuberculosis in East Africa. Fifth Investigation. Tubercle 1970; 51:123-51. 24. Tuberculosis Chemotherapy Centre, Madras. A concurrent comparison of intermittent (twice weekly) isoniazid plus streptomycin and daily isoniazid plus PAS in the domiciliary treatment of pulmonary tuberculosis. Bull WHO 1964; 31:247-71. 25. Miller A, Fox W, ThII R. An international cooperative investigation into thiacetazone (thioacetazone) side-effects. Tubercle 1966; 47:33-73. 26. Colebunders R, Braun M, Nzila N, Dikilu K, Muepu K, Ryder R. Evaluation of the World Health Organisation clinical case definition of AIDS among tuberculosis patients in Kinshasa, Zaire. J Infect Dis 1989; 160:902-3. 27. Styblo K. Overview and epidemiological assessment of the current global tuberculosis situation: with an emphasis on tuberculosis control in developing countries. Bull Int Union Tuberc Lung Dis 1988; 63:39-44.
