Endocarditis

Introduction: Changes in the etiology, epidemiology, and outcome of infective endocarditis (IE) have been observed in recent years. Newer invasive therapeutic interventions have increased the risk of bacteremia and nosocomial endocarditis in the population at risk. A retrospective analysis of hospital-acquired IE cases was performed in a tertiary hospital during 1985 to 1999. Material and methods: Cases included were those classified as "probable" or "definite" by the IE diagnostic criteria of Durack. Nosocomial acquisition was considered if diagnosis was made > 72 h after hospital admission and there was no evidence that IE was present at the time of admission. Patients receiving a diagnosis within 60 days of a previous hospital admission were also classified as nosocomial, when a risk procedure for bacteremia was performed, or when any predisposing factor for IE was present during hospitalization. Early prosthetic valve endocarditis (PVE) cases (< 1 year) were excluded from the analysis. Clinical characteristics, etiology, predisposing cardiac condition, source of infection, and outcome were analyzed. Results were compared with those obtained in community-acquired cases.

Results: Of 493 cases of IE diagnosed over 15 years, 38 were considered to be hospital acquired. Twenty-eight cases were native valve endocarditis (NVE) in non-IV drug user patients, and 10 cases were late PVE. Overall, the most frequent microorganisms involved were staphylococci (58%). The main sources of infection were intravascular procedures or catheter-related infections (55%). When nosocomial NVE cases were compared with community-acquired cases, mortality was greater (29% vs 9.7%) in hospital-acquired endocarditis. Analysis of time trends showed an increased rate of nosocomial cases in NVE throughout the years of the study.

Conclusions: In NVE, the number of cases that are hospital acquired has been increasing during the last 15 years. These cases are frequently associated with invasive intravascular procedures or IV catheter-related infections. Most patients have a previous valvulopathy that predisposes to IE. The spectrum of microorganisms involved is different from the community-acquired cases. Also, the outcome of endocarditis is worse in nosocomial NVE patients.

Key words: native valve; nosocomial endocarditis; prosthetic valve

Abbreviations: [chi]LT = [chi] linear trend; CI = confidence interval; IE = infective endocarditis; NS = not significant; NVE = native valve endocarditis; OR = odds ratio; PVE = prosthetic valve endocarditis

**********

Changes in the etiology, epidemiology, and outcome of infective endocarditis (IE) have been observed, especially in the last 20 years. (1-3) An important aspect of the evolving epidemiology of IE has been an increase in the average age of the patients with this disease. Studies (2,4) have reported a mean age of 55 years in IE patients, with nearly 50% of cases occurring in patients > 60 years old. The incidence of degenerative valve disease (typically seen in the elderly) has increased in parallel with the decrease in the incidence of rheumatic valvulopathy. (3,5) Also, newer invasive therapeutic interventions (particularly IV catheters, pacemakers, and dialysis shunts) have all increased the risk of bacteremia and subsequent endocarditis in the population at risk. (6-8) These changes in the type of acquisition, predisposing cardiac condition for IE, and patient age are also modifying the etiology. In recent years, Staphylococcus aureus has been the most frequent microorganism involved in eases of native valve endocarditis (NVE), in contrast with older series, in which the predominant etiologic agents were viridans group streptococci. (9-12) In this retrospective study, we evaluated the relevance of these changes and the incidence of nosocomial-acquired endocarditis in our institution over the last 15 years.

MATERIALS AND METHODS

We performed a retrospective chart review of all cases of IE diagnosed over a 15-year period from January 1, 1985, to December 31, i999, at the Ramon y Cajal Hospital, a 1,000-bed tertiary referral hospital in Madrid, Spain. Cases were identified by a review of all computerized discharge diagnoses of IE and by searching the consultation files of our Infectious Diseases Service. Cases were classified according to the new diagnostic criteria for IE proposed by Durack et al, (13) and only probable or definite cases of endocarditis were included. Hospital records were examined, and information pertaining to demographic characteristics, clinical, laboratory and echocardiographic features, sites of infection, outcome, and complications were collected.

The acquisition of IE was considered nosocomial if the diagnosis was made > 72 h after hospital admission and if there was no evidence that IE was present at the time of admission. Diagnoses made within 60 days of a previous hospital admission were also classified as nosocomial, or when a risk procedure for bacteremia was performed, or when any predisposing factor for IE was present during hospitalization. Cases of early prosthetic valve endocarditis (PVE) occurring in the first year after implantation as well as pacemaker lead endocarditis were excluded from the analysis.

A predisposing factor for IE was defined as any invasive procedure known to induce bacteremia that had been performed within 60 days prior to IE diagnosis, and the following categories were considered: dental manipulation, gynecology procedures, GI manipulation, urologic manipulation, invasive intravascular techniques (cardiac catheterization, pacemaker insertion, intravascular devices), and previous bacteremia or uncontrolled infection. Cardiac conditions associated with IE were those considered of moderate or high risk for IE included in the recommendations of prevention of bacterial endocarditis elaborated by the American College of Cardiology and the American Heart Association. (14)

Microbiology

Blood samples for culture were drawn under sterile conditions and processed with an automated monitoring system (BACTEC NR660, BACTEC 9240; Becton, Dickinson and Company; Franklin Lakes, NJ). Isolates were identified with standard microbiological tests and automated methods (PASCO; Difco; Detroit, MI) from 1987 to 2000. Before 1987, species were identified using enzymatic or biochemical tests based on the Manual of Clinical Microbiology. (15) Antibiotic susceptibility tests were performed using methods recommended by the National Committee for Clinical Laboratory Standards.

Echocardiography

Patients underwent the transthoracic modality of echocardiographic imaging and transesophagic echocardiography. The echocardiographic studies were done with commercially available instruments (Ultramark 2 and HDI 5000; Advanced Technology Laboratories; Bothell, WA). Two-dimensional, Doppler transthoracic echocardiography was performed with 2- to 4-MHz phase-array transducers. Valvular vegetation was defined as a localized mass of shaggy echoes adherent to a valve leaflet. Nonspecific valvular thickening was not interpreted as vegetation (minor criteria). The vegetative mass was measured in various planes. Size was measured by maximal length and width during freeze-frame analysis. Valvular regurgitation and its severity were graded by color Doppler imaging, using semiquantitative standard criteria. A perivalvular abscess was defined as a circular echo density, without flow in its interior.

Statistical Analysis

The association between categorical variables was performed with the [chi square] test with the Yates correction or the Fisher Exact Test (two-tailed) as needed. Continuous variable association was analyzed with the Mann-Whitney U test. For time trends, analysis for linear proportions was performed with a [chi] linear trend ([chi]LT) test. Statistical analysis was performed using statistical software (SPSS version 9.0; SPSS; Chicago, IL; and EpiInfo 6.0; Centers for Disease Control and Prevention; Atlanta, GA). Statistical significance was defined as p < 0.05.

RESULTS

Four hundred ninety-three cases of IE were diagnosed during 1985 to 1999. Thirty-eight cases were considered hospital acquired. Twenty-eight cases were NVE in non-IV drug user patients, and 10 cases were late PVE (Fig 1).

Nosocomial NVE

One hundred sixty-two cases were NVE in non-IV drug user patients. Twenty-eight cases were acquired nosocomially (17.2%). Detailed characteristics of the cases are shown in Table 1.

Patient Characteristics: The mean age ([+ or -] SD) of the patients was 49.6 [+ or -] 23.2 years (range, 1 to 79 years; median, 50 years). Eighteen patients were male, and 10 were female. Seventy-five percent (21 patients) had a previous cardiac predisposing factor for IE. The most frequent cardiac condition was rheumatic valvular disease (n = 8, 36%). Other cardiac conditions were degenerative valvular disease (n = 2), congenital myocardiopathies (n = 3), obstructive hypertrophic cardiopathy (n = 1), and cardiac valve prosthesis (n = 3).

Microbiology: Coagulase-negative staphylococci were the most frequent microorganisms involved in nosocomial NVE (n = 9 cases, 32%), followed by S aureus (n = 8, 28.5%) and by Enterococcus sp (n = 4, 14.2%). There were some differences in the etiologic agents when nosocomial-acquired cases were compared to those acquired in the community. In community-acquired cases, the most frequent etiologic agents were viridans group streptococci (36%); however, these microorganisms only caused 4% of the hospital-acquired NVE cases (p < 0.001; odds ratio [OR], 0.10; 95% confidence interval [CI], 0.01 to 0.69). Coagulase-negative staphylococci were the most frequent microorganisms in nosocomial eases (32%), although they were an uncommon etiology in community-acquired cases (4%) [OR, 12.4; 95% CI, 3.2 to 48.6; p < 0.001]. NVE cases due to fungal etiology were all nosocomially acquired, and no cases were found in the community-acquired NVE group (p = 0.005) [Table 2].

Source of Infection: A predisposing risk factor for IE was present in 24 cases of the 38 nosocomial cases (88.5%), but in only 37 of the 134 of the community-acquired NVE cases (27.6%) [OR, 15; 95% CI, 5 to 48; p < 0.001]. The most frequent sources of infection in nosocomial cases were invasive IV procedures (cardiac catherization [n = 6], pacemaker implantation [n = 1]), and IV catheter-related infection (n = 8, 32%). Two other patients underwent cardiac catheterization and cardiac surgery. Two patients had skin diseases that may predispose to skin-related bacteremia (psoriasiform dermatitis with skin biopsy and Lyell syndrome, one ease each). Genitourinary tract procedures were the source of infection in two cases (delivery in one case, and urinary catheterization in other).

Cardiac Involvement: Twenty-three cases (82%) were left-sided endocarditis. The valves involved were the mitral valve (50%), aortic valve (11%), tricuspid valve (11%), aortic and mitral valves (11%), and mural endocarditis (7%).

Outcome: Nineteen cases (68%) had a complication associated with endocarditis. Systemic embolisms occurred in 36% of the nosocomial NVE cases. Six cases had neurologic events (21%). Fifty percent (14 cases) underwent surgical therapy. The frequency of mortality in patients with nosocomial NVE was 29% (n = 8). The frequency of mortality in patients with community-acquired NVE was 10% (n = 13) [OR, 3.7; 95% CI, 1.4 to 10; p = 0.012]. Etiologies of the eight cases with a fatal outcome were caused by S aureus (n = 3), Staphylococcus epidermidis (n = 4), and polymicrobial etiology (n = 1). Two of these patients died due to causes not related to endocarditis.

Time Trend: Analysis of the time trends of the frequency of nosocomial NVE eases revealed that the number of eases with hospital acquisition has increased during the 15 years of the study. During the period from 1985 to 1987, the frequency of nosocomial NVE was 3.4% of the NVE eases diagnosed. The frequency of nosocomial NVE during the last period analyzed, 1997 to 1999, was 31% (nine cases). The change in the number of NVE eases nosocomially acquired was statistically significant when linear trend of proportions was used ([chi]LT = 9.7; p = 0.001) [Fig 2].

Nosocomial Late PVE

One hundred cases were PVE. Forty-seven cases were early PVE (< 12 months after surgery), and 53 cases were late PVE (> 12 months). Early PVE cases were considered by definition to be nosocomial acquired. In the late PVE cases, 19% were acquired nosocomially (10 cases).

Patient Characteristics: Five patients were male, and five were female. Mean age was 43 years (range, 1 to 74 years). Mean time from valve prosthesis replacement to endocarditis diagnosis was 3,094 days (range, 880 to 5,400 days; median, 3,172 days).

Microbiology: The most frequent etiologic agents were staphylococci causing 50% of the cases: coagulase-negative staphylococci (three cases), followed by S aureus (two cases). Fungal etiology caused two cases. Microorganisms causing endocarditis were similar in patients with late hospital-acquired PVE and community-acquired PVE, but cases caused by Candida sp were only nosocomial (p = 0.032) [Table 3].

Source of Infection: A source of infection or a predisposing risk factor for PVE were identified in half of the nosocomial cases; this circumstance was significantly less frequent in the late community-acquired cases (p = 0.036; OR, 5.1; 95% CI, 1 to 29). Two patients had catheter-related infections, and one patient had an aortic valve replacement (and mitral prosthesis endocarditis developed).

Cardiac Involvement: Seven eases affected mitral prosthesis (70%), and there was also one case each of aortic, aortic and mitral, and patch involvement.

Outcome: Two patients underwent valve surgery (20%). Forty percent of cases had no complications associated with endocarditis. The mortality rate of the nosocomial late PVE cases was 20%, and the mortality rate of the community-acquired late PVE cases was 14% (not significant [NS], p = 0.63).

Time Trend: In the analysis of the time trend of late PVE according to the type of acquisition, the frequency of hospital-acquired cases did not change significantly. Analysis for linear proportions ([chi]LT) could not be calculated because the expected frequency was < 1 in one of the period of time considered (1985 to 1987) [Fig 3].

DISCUSSION

Since the first description of IE published by William Osler in 1885, this infectious disease has experienced important changes, mainly during the last 20 years. (1-3,10) These changes are modifying the epidemiology, etiology, diagnostic methods, and treatment.

The major changes in the epidemiology of IE in developed countries include the older age of the patients, the occurrence of new predisposing factors for IE, and the increased rate of hospital-acquired cases, which is especially associated with the more frequent use of intravascular devices and intravascular procedures. (1,4,6) These changes are associated with a concomitant switch in the etiologic microorganisms that cause IE. In some series of NVE, (5,9-11) S aureus is more frequent than viridans group streptococci, the main group of microorganisms causing IE during the 1970s and 1980s.

Hospital-acquired endocarditis reported in other series (6,10,16,17) represent 10 to 30% of all the IE cases. Ferrnandez-Guerrero et al (6) reported a rate of 9.3% of nosocomial endocarditis not related to cardiac surgery in their series, which included 248 episodes of endocarditis diagnosed from 1978 to 1992. They found a 10-fold increase in the number of hospital-acquired cases diagnosed during 1980s compared with their historical series. (6)

Although the proportion of hospital-acquired cases in our study is similar to that found in other series, we have observed that the frequency of nosocomial acquired cases has experienced an increase during the 15 years of the study. In the first period of the study (1985 to 1987), nosocomial NVE cases represented 3.4% of the NVE cases diagnosed during these years, but the frequency of hospital acquired cases reached 31% of the NVE cases in the last period (1997 to 1999). This change in the number of nosocomial NVE cases was statistically significant ([chi]LT = 9.7; p < 0,01).

This finding could be explained by the increase in the number of nosocomial bloodstream infections, the number of patients admitted to ICUs, the increase in the number of invasive vascular procedures, and the higher incidence of degenerative valve disease in our aging population. Although in the study of Fernandez-Guerrero et al (6) only 23% of nosocomial cases had previous valvulopathy, and 32% in the series of Lamas and Eykyn, (18) in our series 72% of nosocomial-acquired NVE cases had any predisposing cardiac condition; rheumatic disease was the most frequent heart condition (36%). Although rheumatic fever is considered eradicated in our country, due to the high incidence of rheumatic fever in those previous decades (until the 1960s), we now have a population aged 50 to 80 years that suffered valvular disease as a sequelae of the episode in the childhood, and rheumatic valvulopathy is still the most frequent underlying heart disease in our study. However, it should be noted that this might not be the ease in another context. Chronic rheumatic disease is now rare in industrialized countries, although this group of patients has been replaced by new at-risk groups including elderly people with degenerative valve disease, patient with intravascular prostheses, and patients with intravascular devices. (19)

The source of infection determines the profile of microorganisms causing endocarditis. When the source was intravascular devices or procedures, staphylococci were the most frequent etiologic agents; when the source was the genitourinary tract, Enterococcus sp was the predominant microorganism. (6) Fernandez-Guerrero et al (6) found that the source of infection was IV catheterization in 56.5% of cases, genitourinary tract instrumentation in 30%, total parenteral nutrition in 8.6%, and with liver biopsy in 4.3%. In our study, the most frequent predisposing factors were IV catheter-related infections (32%) and intravascular invasive procedures (25%).

The first cases of endocarditis due to S aureus associated with IV catheters were reported by Watanakunakorn and Baird. (20) After this article, other investigators (6,9,12,17) reported cases of IE related to intravascular catheters, with S aureus being the most frequent microorganism. Fowler et al (21) reported that intravascular devices were the most common source of infection (50.8%) in 59 eases of endocarditis caused by S aureus diagnosed in their institution from 1994 to 1998. Half of them were long-term venous access or hemodialysis catheters, and 14% were peripheral catheters.

The risk of endocarditis developing after an episode of S aureus bacteremia in patients with intravascular devices remains controversial. In a prospective study by Fowler et al (22) in this group of patients, 23% had endocarditis evident on transesophagic echocardiography. In a study published by Chang et al, (9) the rate of endocarditis due to S aureus with intravascular devices as source of bacteremia was 11%. The need to perform echocardiography in all patients with bacteremia due to S aureus is debated by some authors. Pigrau et al (23) considered that echocardiography did not need to be performed in patients with uncomplicated bacteremia (no predisposing valvulopathy, persistent fever, metastatic loci).

Other series (6,8,18) of nosocomial endocarditis reported the frequent association with intravascular devices and S aureus as the most frequent etiologic agent. Gouello et al (8) reported 22 hospital-acquired endocarditis cases occurred in an ICU. Forty-one percent of these patients had previous valvulopathy. The source of infection was identified in 54.5% of cases, and 90% of these cases were associated with intravascular devices. S aureus was the most frequent microorganism (77.2%). (8) Lamas and Eykyn (18) reported 22 nosocomial NVE cases diagnosed over 11 years (1985 to 1996). The incidence of hospital-acquired cases was 14%. Seventy-seven percent were caused by staphylococci, and the main source of infection was an intravascular device. (18)

In an article by Munoz et al, (24) aimed at defining the incidence, risk factors, and characteristics of the bacteremias occurring during the first 72 h after nonsurgical cardiologic invasive procedures, the incidence of bacteremia was 0.11% after these procedures. Only 16% of the patients with bacteremia had a previous valvulopathy, and none had endocarditis. (24) In our study, all the cases related to cardiac catheterization were in patients with previous valvulopathy, and endocarditis was always left sided. The diagnosis of endocarditis was performed a mean of 28 days after cardiac catheterization.

In our study, the most frequent etiologic agent causing nosocomial endocarditis was coagulase-negative staphylococci, in 32% of the cases. In other series, (6,16,18,25) the frequency of nosocomial endocarditis caused by coagulase-negative staphylococci was lower (1 to 7%).

The frequency of mortality of nosocomial-acquired eases in our study was 29% higher than the mortality frequency of community-acquired-acquired NVE cases (9.7%). This high mortality rate has been reported also by other authors. Fernandez-Guerrero et al (6) reported a mortality rate of 48% in their nosocomial cases. Gouello et al (8) had a mortality rate directly related with nosocomial endocarditis of 36%. Lamas and Eykyn et al (18) reported a mortality rate of 50% in patients with nosocomial NVE. In a series by Mouly et al, (10) analyzing the mortality of patients with IE diagnosed in their institution, the hospital acquisition was identified as an independent risk factor associated with mortality.

In conclusion, the number of hospital-acquired cases of NVE has increased during the last years. These cases are frequently associated with invasive intravascular procedures or IV catheter-related infection. Most patients have a previous valvulopathy that predisposes to IE. The spectrum of microorganisms involved is different than that of the community-acquired cases. These data should be considered when empiric therapy for suspected nosocomial IE is initiated. Noteworthy, the outcome of endocarditis is worse in patients with nosocomial NVE. Physicians must be aware of the risk of endocarditis developing during hospitalization in patients who undergo invasive procedures that potentially cause bacteremia. Patients with a previous valvulopathy subjected to these procedures may benefit from anticipated or prophylactic antimicrobial therapy; and in any case, prompt echocardiography and empiric therapy are compelling if patients become febrile or bacteremic.

REFERENCES

(1) Cabell CH, Jollis JG, Peterson GE, et al. Changing patient characteristics and the effect on mortality in endocarditis. Arch Intern Med 2002; 162:90-94

(2) Hoen B, Alla F, Selton-Suty C, et al. Changing profile of infective endocarditis: results of a 1-year survey in France. JAMA 2002; 288:75-81

(3) Fefer P, Raveh D, Rudensky B, et al. Changing epidemiology of infective endocarditis: a retrospective survey of 108 cases, 1990-1999. Eur J Clin Microbiol Infect Dis 2002; 21:432-437

(4) Tornos MP, Olona M, Permanyer-Miralda G, et al. Is the clinical spectrum and prognosis of native valve infective endocarditis in non-addicts changing? Eur Heart J 1995; 16:1686-1691

(5) Watanakunakorn C, Burkert T. Infective endocarditis at a large community teaching hospital, 1980-1990: a review of 210 episodes. Medicine (Baltimore) 1993; 72:90-102

(6) Fernandez-Guerrero ML, Verdejo C, Azofra J, et al. Hospital-acquired infectious endocarditis not associated with cardiac surgery: an emerging problem. Clin Infect Dis 1995; 20:16-23

(7) Gilleece A, Fenelon L. Nosocomial infective endocarditis. J Hosp Infect 2000; 46:83-88

(8) Gouello JP, Asfar P, Brenet O, et al. Nosocomial endocarditis in the intensive care unit: an analysis of 22 cases. Crit Care Med 2000; 28:377-382

(9) Chang FY, MacDonald BB, Peacock JE Jr, et al. A prospective multicenter study of Staphylococcus aureus bacteremia: incidence of endocarditis, risk factors for mortality, and clinical impact of methicillin resistance. Medicine (Baltimore) 2003; 82:322-332

(10) Mouly S, Ruimy R, Launay O, et al. The changing clinical aspects of infective endocarditis: descriptive review of 90 episodes in a French teaching hospital and risk factors for death. J Infect 2002; 45:246-256

(11) Espersen F, Frimodt-Moller N. Staphylococcus aureus endocarditis: review of 119 cases. Arch Intern Med 1986; 146: 1118-1121

(12) Watanakunakorn C. Staphylococcus aureus endocarditis at a community teaching hospital, 1980 to 1991: an analysis of 106 cases. Arch Intern Med 1994; 154:2330-2335

(13) Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med 1994; 96:200-209

(14) Dajani AS, Taubert KA, Wilson W, et al. Prevention of bacterial endocarditis: recommendations by the American Heart Association. Circulation 1997; 96:358-366

(15) Lennette EH, Balows A, Hausler WJ Jr, et al, eds. Manual of clinical microbiology. 4th ed. Washington, DC: American Society for Microbiology, 1985

(16) Friedland GH, von Reyn CF, Levy B, et al. Nosocomial endocarditis. Infect Control 1984; 5:284-288

(17) Terpenning MS, Buggy BP, Kauffman CA. Hospital-acquired infective endocarditis. Arch Intern Med 1988; 148:1601-1603

(18) Lamas CC, Eykyn SJ. Hospital acquired native valve endocarditis: analysis of 22 cases presenting over 11 years. Heart 1998; 79:442-447

(19) Michel L, Acar J. Native cardiac disease predisposing to infective endocarditis. Eur Heart J 1995; 16(suppl B):2-6

(20) Watanakunakorn C, Baird IM. Staphylococcus aureus bacteremia and endocarditis associated with a removable infected intravenous device. Am J Med 1977; 63:253-256

(21) Fowler VG Jr, Sanders LL, Kong LK, et al. Infective endocarditis due to Staphylococcus aureus: 59 prospectively identified cases with follow-up. Clin Infect Dis 1999; 28:106-114

(22) Fowler VG Jr, Li J, Corey GB, et al. Role of echocardiography in evaluation of patients with Staphylococcus aureus bacteremia: experience in 103 patients. J Am Coll Cardiol 1997; 30:1072-1078

(23) Pigrau C, Rodriguez D, Planes AM, et al. Management of catheter-related Staphylococcus aureus bacteremia: when may sonographic study be unnecessary? Eur J Clin Microbiol Infect Dis 2003; 22:713-719

(24) Munoz P, Blanco JR, Rodriguez-Creixems M, et al. Bloodstream infections after invasive nonsurgical cardiologic procedures. Arch Intern Med 2001; 161:2110-2115

(25) Chen SC, Dwyer DE, Sorrell TC. A comparison of hospital and community-acquired infective endocarditis. Am J Cardiol 1992; 70:1449-1452

* From the Departments of Infectious Diseases (Drs. Martin-Davila, Fortun, Navas, Cobo, and Moreno) and Cardiology (Drs. Jiminez-Mena and Moya), Hospital Ramon y Cajal, Madrid, Spain.

Manuscript received January 12, 2005; revision accepted February 10, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Pilar Martin-Davila, MD, PhD, Servicio de Enfermedades Infecciosas, Hospital Ramon y Cajal, Ctra. Colmenar km. 9,500, 28034 Madrid, Spain; e-mail: pmartin.hrc@salud. madrid.org

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group