Pneumococcal pneumonia in HIV-infected patients by antiretroviral therapy periods

Transcription

1 DOI: /j x r 2008 British HIV Association HIV Medicine (2008), 9, ORIGINAL RESEARCH Pneumococcal pneumonia in HIV-infected patients by antiretroviral therapy periods M Saindou, 1 C Chidiac, 2 P Miailhes, 3 N Voirin, 1,4 D Baratin, 4 M Amiri, 3 J-M Livrozet, 5 J-L Touraine, 5 C Trepo, 3 D Peyramond 6 and P Vanhems 1,4 1 Université de Lyon; Université Lyon 1; CNRS, UMR 5558, Laboratoire de Biométrie et Biologie Evolutive, Epidémiologie et Santé Publique, 8 avenue Rockefeller, Lyon F-69373, France, 2 Université Lyon 1, Lyon, France; Faculté Laennec INSERM, U851, Lyon, France; Hospices Civils de Lyon, Hôpital de la Croix-Rousse, Service des Maladies Infectieuses et Tropicales, Lyon cedex 04, France, 3 INSERM, U871, Lyon, France; Université Lyon 1, IFR62 Lyon-Est, Lyon, France; Hospices Civils de Lyon, Hôpital Hôtel-Dieu, Service d Hépato-gastroentérologie, Lyon, France, 4 Hospices Civils de Lyon, Hôpital Edouard Herriot, Département d Hygiène, Epidémiologie et Prévention, 5 place d Arsonval, Lyon, F-69437, France, 5 Hospices Civils de Lyon, Hôpital Edouard Herriot, Service d Immunologie Clinique, 5 place d Arsonval, Lyon, F-69437, France and 6 Hospices Civils de Lyon, Hôpital de la Croix-Rousse, Service des Maladies Infectieuses et Tropicales, Lyon cedex 04, France Objective To ascertain the relationship between periods of various antiretroviral therapies and the incidence of first community-acquired pneumococcal pneumonia (CAPP) among HIV-1 infected patients. Methods We analysed 4075 patients enrolled prospectively in the Lyon section of the French Hospital Database on HIV between 1993 and 2004, stratified into three groups. The first group (G1) included patients for whom enrolment and last follow-up were before the highly active antiretroviral therapy (HAART) period (beginning 1 July 1996); the second group (G2) comprised patients who were enrolled before HAART but had last follow-up in the HAART period; the third group (G3) included patients for whom both enrolment and last follow-up took place in the HAART period. Results Fifty-five CAPP episodes were identified. The incidence of CAPP per 1000 patient-years declined over time, from 10.6 to 1.5 and 2.5 in calendar periods G1, G2 and G3, respectively (P for linear trend). Factors associated with a decreased risk of CAPP were lower age, baseline CD4 count 200 cells/ml and more recent years of enrolment, when HAART use became extensive (Po0.001). The use of intravenous drugs increased the risk of CAPP (Po0.001). Conclusions There has been a significant reduction in the incidence of CAPP in HIV-1 infected patients since the advent of HAART. Keywords: acquired immune deficiency syndrome, community-acquired pneumococcal pneumonia, highly active antiretroviral therapy, human immunodeficiency virus, Lyon Received: 22 November 2007, accepted 20 December 2007 Introduction Correspondence: Philippe Vanhems, Université de Lyon, Université Lyon 1, CNRS, UMR 5558, Laboratoire de Biométrie et Biologie Evolutive, Epidémiologie et Santé Publique, 8 Avenue Rockefeller, Lyon F-69373, France. Tel: ; fax: ; philippe.vanhems@chu-lyon.fr Morbidity and mortality have been reduced dramatically in HIV-1 infected individuals because of partial immune restoration with highly active antiretroviral therapies (HAART) [1]. Preventive chemotherapy has also contributed to the decreased incidence of opportunistic complications [2 4]. Streptococcus pneumoniae is the main potential cause of community-acquired pneumonia in both the general population and HIV-1 infected subjects [5]. Few results have been reported on trends in the incidence rate of community-acquired pneumococcal pneumonia (CAPP) since the advent of HAART. In the present study, we present data on a prospective cohort of 203

2 204 M Saindou et al. hospitalized individuals to profile the incidence of CAPP infection among HIV-1 subjects before and after the introduction of HAART. The objective is to compare risk factors of CAPP among HIV-1 patients by time periods corresponding to different eras of antiretroviral therapy (ART). Patients and methods The French Hospital Database on HIV is a clinical, epidemiological network started in 1989 for medical centres involved in the medical care of HIV-infected persons at French hospitals. National coverage, achieved in 1992, is reported in detail elsewhere [6]. This study was undertaken among 4075 patients enrolled in Lyon University hospitals between 1993 and A full cohort description can be found elsewhere [7]. Briefly, patients with a documented HIV infection who had signed an informed consent form were enrolled at their first hospital visit. A standardized data collection form was filled out, at enrolment and at least every 6 months. Demographic, clinical, biological and treatment data were recorded regularly, and clinical events were based on standardized definitions [6]. Cases of CAPP were reported through a database surveillance system operating in Lyon University hospitals. CAPP was considered if patients presented with fever and respiratory symptoms and had radiological findings consistent with pneumonia associated with bacteriological confirmation of pneumococcus as causative agent and/or sum clinical and radiological findings in favour of pneumococcal pneumonia. Among the 55 patients with CAPP, diagnosis was based on bronchoalveolar lavage (BAL) (n 5 30), blood culture (n 5 14), sputum (n 5 5) or urinary antigen (n 5 2). For four cases, the diagnosis was highly suspected based on chest X-ray, clinical presentation and favourable outcome after the use of antibiotics targeted on pneumonia. Other cases of pulmonary infection or pulmonary diseases (e.g. cardiovascular conditions or cancer) were excluded. Only initial cases of bacterial pneumonia were considered. Calendar years were stratified into two periods: before 1 July 1996 and after 1 July Respectively, these periods represent the pre-haart and HAART eras, consisting of a combination of at least three antiretroviral drugs. Patients were stratified into three groups according to the date of study entry and the last date of follow-up in the cohort by antiretroviral therapy periods. The first group (G1) included patients for whom enrolment and last follow-up took place before 1 July 1996; the second group (G2) contained patients for whom enrolment took place before 1 July 1996 and last follow-up occurred during the HAART period (after 1 July 1996); the third group (G3) comprised patients for whom both enrolment and last follow-up took place during the HAART period The w 2 test or Fisher s exact test was employed to compare categorical variables. The incidence of CAPP was calculated per 1000 patient-years of follow-up and was compared inside each calendar period. Cox s proportional hazards model was implemented to calculate the adjusted hazards ratio (HR), with its 95% confidence interval (95% CI) of CAPP over calendar periods. Variables were included in the multivariate model according to a backward procedure based on the likelihood ratio test. The statistically significant threshold was considered to be Po0.05. Results Of the 4075 patients included in the study, 3192 (78.3%) were male and 883 (21.7%) were female (Table 1). Among the 55 CAPP cases, 47 (85.5%) were male and eight (14.5%) were female. Mean age at baseline was 36.0 years (standard deviation 10.2). Mean duration of follow-up was 5.3 years: 2.4 years for G1, 9.9 years for G2 and 3.6 years for G3. Mean CD4 counts were 372 cells/ml at baseline, 298 cells/ml at G1, 410 cells/ml at G2 and 374 cells/ml at G3. Overall, 21.7% of patients in G1 received antiretroviral therapy as mono or bi-therapy: 27.0% in G2 and 13.4% in G3. The proportion of patients who received HAART was 30.0% in G3, 1.3% in G2 and 0.0% in G1 (Po0.001). Overall, 55 CAPP episodes were identified: 17 in G1, 18 in G2 and 20 in G3. The attack rate of CAPP for 100 patients differed significantly by time (P ) (Table 2). The incidence of CAPP per 1000 patient-years decreased over time, from 10.6 (95% CI ) to 1.5 (95% CI ) and 2.5 (95% CI ) in calendar periods G1, G2 and G3, respectively (P for linear trend) (Table 2). By multivariate analysis (Table 3), lower age, baseline CD4 count 200 cells/ml and more recent year of entry remained independently associated with a reduced risk of CAPP. The use of intravenous drugs increased the risk of CAPP (Po0.001). Discussion We observed a significant decline in CAPP incidence among HIV-infected persons after the advent of HAART. The significant relationship of CAPP incidence to time confirmed the clinical benefit of HAART. It is reasonable to hypothesize that this result was largely the consequence of the well-known immune restoration induced by HAART [1]. However, other factors should also be considered as explanations for the reduced incidence of CAPP. Knowledge of HIV could be advanced and quality of care could be

3 Pneumococcal pneumonia in HIV-infected patients 205 Table 1 Baseline characteristics of HIV-1 patients by calendar time period groups in Lyon University Hospitals Inclusion trend of hospital database patient groups Patient characteristics Group 1* Group 2 w,z Group 3 Total n (%) n (%) n (%) n (%) Gender Female 111 (16.6) 259 (20.9) 513 (23.6) 883 (21.7) Male 556 (83.4) 980 (79.1) 1656 (76.4) 3192 (78.3) Age group (years) o (30.3) 470 (37.9) 581 (26.8) 1253 (30.8) (61.2) 679 (54.8) 1307 (60.2) 2394 (58.7) (8.5) 90 (7.3) 281 (13.0) 428 (10.5) Presumed route of HIV infection Heterosexual intercourse 392 (58.8) 644 (52.0) 935 (43.1) 1971 (48.4) Bisexual/homosexual intercourse 151 (22.6) 379 (30.6) 840 (38.7) 1370 (33.6) Intravenous drug use 98 (14.7) 141 (11.4) 108 (5.0) 347 (8.5) Other/unknown 26 (3.9) 75 (6.0) 286 (13.2) 387 (9.5) CD4 lymphocyte counts category (cells/ml) z o (40.0) 274 (22.1) 572 (26.4) 1113 (27.3) (52.0) 846 (68.3) 1481 (68.3) 2674 (65.6) CDC disease stage A 286 (42.9) 820 (66.2) 1510 (69.6) 2616 (64.2) B 161 (24.1) 281 (22.7) 332 (15.3) 774 (19.0) C 220 (33.0) 138 (11.1) 327 (15.1) 685 (16.8) Pneumococcus pneumoniae during follow-up Yes 17 (2.5) 18 (1.5) 20 (0.9) 55 (1.3) *Includes patients for whom entry and last follow-up took place before 1 July w Includes patients for whom entry took place before 1 July 1996 and last follow-up took place during the highly active antiretroviral therapy (HAART) period z Pneumococcal pneumonia (CAPP) occurred in this group during the HAART period. Includes patients for whom entry and last follow-up took place during the HAART period z In some cases the total did not reach 100% because of missing data. Table 2 Community-acquired pneumococcal pneumonia incidence and attack rate in three periods with point estimates and 95% confidence intervals (CI) Inclusion period Number of patients Follow-up in cumulative years Number of pneumonia patients Attack rate per 100 patient-years Incidence rate per 1000 patient-years (95% CI) Group 1* ( ) Group 2 w,z ( ) Group ( ) *Includes patients for whom entry and last follow-up took place before 1 July w Includes patients for whom entry took place before 1 July 1996 and last follow-up took place during the highly active antiretroviral therapy (HAART) period z Pneumococcal pneumonia (CAPP) occurred in this group during the HAART period. Includes patients for whom entry and last follow-up took place during the HAART period P-value of CAPP: attack rate (P ) and incidence rate (P trend ). improved. We report a regular drop in the attack rate of CAPP since the introduction of HAART. Previous studies have suggested that HAART is protective against pneumonia [8], with most AIDS-related opportunistic infections decreasing in incidence in the HAART era. The occurrence of CAPP was associated with host risk factors, such as age group years, intravenous drug use as the route of HIV infection and CD4 count o200 cells/ml. Similar trends have been reported in 2005 [2]. A positive point of this prospective study was that hospital-based populations with HIV infection provided opportunities to follow the natural history of HIV and related opportunistic infections. Access to care is an important issue in pneumonia incidence because of the previously reported clinical benefits of preventing opportunistic infections by antiretroviral therapy and other prophylaxes. Patient recruitment and the diagnostic approach to CAPP infection have not changed with time period. Our study has some limitations. The cigarette smoking variable and the characteristics of HAART regimens and treatment interruptions were not analysed in detail. We

4 206 M Saindou et al. Table 3 Univariate and multivariate analyses of risk factors for community-acquired pneumococcal pneumonia among HIV-1 patients in Lyon University Hospitals by calendar time periods Cox models Pneumonia status Univariate analysis Multivariate analysis Baseline characteristics Positive, no. (%) n 5 55 Negative, no. (%) n HR (95% CI) P HR (95% CI) P Gender Female 8 (14.5) 875 (21.8) 1.0 Male 47 (85.5) 3145 (78.2) 1.5 ( ) Age group (years) o30 9 (16.4) 1244 (30.9) (70.9) 2355 (58.6) 2.8 ( ) ( ) (12.7) 421 (10.5) 3.1 ( ) ( ) Presumed route of HIV infection Heterosexual intercourse 25 (45.5) 1946 (48.4) Bisexual/homosexual intercourse 17 (30.9) 1353 (33.6) 1.1 ( ) ( ) Intravenous drug use 11 (20.0) 336 (8.4) 2.5 ( ) ( ) Other/unknown 2 (3.6) 385 (9.6) 0.5 ( ) ( ) CD4 lymphocyte counts category (cells/ml)* o (47.3) 1087 (27.0) (36.4) 2654 (66.0) 0.2 ( ) o ( ) o0.001 CDC disease stage A 22 (40.0) 2594 (64.5) 1.0 B 15 (27.3) 759 (18.9) 2.3 ( ) C 18 (32.7) 667 (16.6) 4.3 ( ) o0.001 Patient group Group 1 w 17 (30.9) 650 (16.2) Group 2 z, 18 (32.7) 1221 (30.4) 0.07 ( ) o ( ) o0.001 Group 3 z 20 (36.4) 2149 (53.4) 0.2 ( ) o ( ) CI, confidence interval; HR, hazards ratio. *In some cases the total did not reach 100% because of missing data. w Includes patients for whom entry and last follow-up took place before 1 July z Includes patients for whom entry took place before 1 July 1996 and last follow-up took place during the highly active antiretroviral therapy (HAART) period Pneumococcal pneumonia (CAPP) occurred in this group during the HAART period. z Includes patients for whom entry and last follow-up took place during the HAART period found that only 30% of patients in G3 received HAART. Such a finding is related to the fact that at baseline visit, HAART is proposed to relevant patients but the effective start of HAART might be delayed by a couple of weeks. We did not collect information on pneumonia vaccination; trials have revealed the positive effect of pneumococcal vaccination [9]. Survival bias might have occurred because of longer survival after the introduction of HAART. However, longer survival allows infectious events to be tracked. We only analysed CD4 cell count at baseline and not the slope of CD4 cell count during the entire follow-up period. However, baseline CD4 cell count remained a good predictor of HIV disease progression [10]. HIV viral load at baseline and HIV viral load trends were not assessed because they were only recorded from 1992 in the database. The likelihood of a change in the means of diagnosis by time seems to be low. However, among recent-entry patients, the diagnosis in those with pulmonary symptoms could be made earlier than in the first years of the epidemic. This bias could affect the severity or clinical presentation of CAPP more than its incidence. Conclusions We observed a significant reduction in the incidence rate of CAPP in HIV-1 infected patients after the introduction of HAART. Despite the declines in CAPP incidence with HAART, it continues to cause morbidity and mortality among HIV-infected subjects. Further research is needed to clarify the effects of other factors in CAPP prevention among patients with HIV, such as pneumococcal vaccination, penicillin and macrolide resistance in the HAART era. Acknowledgements The authors are indebted to C Delorme, E Garcia, MA Nerand and the following collaborators who provided data access: F Bailly, P Barlet, F Biron, F Bissuel, A Boibieux, J-P

5 Pneumococcal pneumonia in HIV-infected patients 207 Bonnin, D Bouhour, F Brunel-Dalmas, E Carbonnel, P Chiarelo, P Chossegros, A Claudy, M Faure, V Gueripel, F Jeanblanc, J-J Jourdan, T Saint-Marc, D Makloufi, L Cotte, C Pariset, S Radenne, G Retornaz, P Rougier, I Schlienger and L Thomas. They thank P André, P Chevallier-Queyron and J Ritter from the virology laboratory of the Lyon University Hospital, and also thank Mr O Da Silva for editing the manuscript. References 1 Mocroft A, Ledergerber B, Katlama C et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet 2003; 362: Heffeman RT, Barrett NL, Gallagher KM et al. Declining incidence of invasive Streptococcus pneumoniae infections among persons with AIDS in the era of highly active antiretroviral therapy, JInfectDis2005; 191: Navin TR, Rimland D, Lennox JL et al. Risk factors for community-acquired pneumonia among persons infected with human immunodeficiency virus. J Infect Dis 2000; 181: Le Moing V, Rabaud C, Journot V et al. Incidence and risk factors of bacterial pneumonia requiring hospitalization in HIV-infected patients started on a protease inhibitorcontaining regimen. HIV Med 2006; 7: Rimland D, Navin TR, Lennox JL et al. Prospective study of etiologic agents of community-acquired pneumonia in patients with HIV infection. AIDS 2002; 16: Tassie J-M, Grabar S, Lancar R, Deloumeaux J, Bentata M, Costagliola D. Time to AIDS 1992 to 1999 in HIV-1-infected subjects with known date of infection. J AIDS 2002; 30: Gayet-Ageron A, Baratin D, Marceillac E et al. The AIDS epidemic in Lyon: patient characteristics and defining illnesses between 1985 and HIV Med 2004; 5: Sullivan JH, Moore RD, Keruly JC, Chaisson RE. Effect of antiretroviral therapy on the incidence of bacterial pneumonia in patients with advanced HIV infection. Am J Respir Crit Care Med 2000; 162: Feikin DR, Feldman C, Schuchat A, Janoff EN. Global strategies to prevent bacterial pneumonia in adults with HIV disease. Lancet Infect Dis 2004; 4: Miller V, Staszewski S, Sabin C et al. CD4 lymphocyte count as a predictor of the duration of highly active antiretroviral therapy-induced suppression of human immunodeficiency virus load. J Infect Dis 1999; 180: