The Impact of Blood Cultures on Antibiotic Therapy in Pneumococcal Pneumonia

The Impact of Blood Cultures on Antibiotic Therapy in Pneumococcal Pneumonia* Grant W. Waterer, MBBS; S. Gregory Jennings, MD; and Richard G. Wunderink, MD, FCCP

Introduction: The cost-effectiveness of blood cultures in community-acquired pneumonia (CAP) has been questioned. Although penicillin-resistant Streptococcus pneumoniae is an increasing problem, penicillin therapy, where appropriate, reduces cost and may reduce antibiotic resistance. Blood cultures, however, can only reduce cost if physicians are prepared to alter therapy based on the results. We reviewed our experience to determine how often physicians changed management based on blood culture results positive for S pneumoniae. Methods: Retrospective chart review was performed of all CAP admissions between January 1996 and December 1998 with blood culture results positive for S pneumoniae. Results: Seventy-four patients out of 1,805 patients admitted with CAP during this period had pneumococcemia. Penicillin resistance was identified in 15 cases (20.3%; high grade in 4 cases) with cephalosporin resistance in 4 of these cases (1 high grade). Fifty-one patients had initial empiric therapy with a third-generation cephalosporin, and 58 patients had empiric coverage of atypical organisms; no patient received empiric penicillin therapy. Blood culture results altered management in 31 patients (41.9%), but in only 2 cases was this due to antibiotic resistance. Fifty-one patients without penicillin allergy grew penicillin-sensitive pneumococci; only 11 patients (21.6%) were changed to penicillin therapy. Thirteen of 35 patients (37.1%) who were given an additional antibiotic for atypical coverage had this antibiotic ceased. Conclusion: Despite evidence of penicillin-sensitive pneumococcal CAP, physicians were reluctant to narrow antibiotic therapy, potentially adding to treatment cost and reducing the impact of blood culture results on management. The impact of penicillin resistance was reduced by the usual empiric choice of a third-generation cephalosporin. While positive blood culture results can clearly be useful in the management of patients with CAP, their cost-effectiveness needs to be assessed in prospective clinical trials. (CHEST 1999; 116:1278 –1281) Key words: antibiotics; bacteremia; community-acquired; penicillin; pneumonia Abbreviations: APACHE 5 acute physiology and chronic health evaluation; CAP 5 community-acquired pneumonia; MIC 5 minimum inhibitory concentration; PNSP 5 penicillin-nonsusceptible Streptococcus pneumoniae; PRSP 5 penicillin-resistant Streptococcus pneumoniae; PSI 5 Pneumonia Severity Index

patients with community-acquired pneumonia I n(CAP), blood cultures are frequently performed as an aid to the diagnosis of the causative organism. The detection of an organism resistant to the empiric antibiotic therapy chosen is an important outcome of blood cultures. However, with the current trend toward broad-spectrum empiric therapy, the most *From Methodist Healthcare, Memphis, TN. Dr. Waterer has been supported by a grant from the MethodistLeBonheur Healthcare Foundation and by the Athelstan and Amy Saw Medical Research Fellowship from the University of Western Australia. Manuscript received January 14, 1999; revision accepted May 20, 1999. Correspondence to: Grant W. Waterer, MBBS, Pulmonary Host Defense Fellow, Methodist Healthcare, 1265 Union Ave, 501 Crews Wing, Memphis, TN 38104-2499; e-mail: waterer@ ibm.net 1278

likely impact on management is the possible reduction or narrowing of antibiotic therapy. Due to the overall low yield of blood cultures in CAP, several recent reviews have questioned their cost-effectiveness in the management of CAP.1–3 Streptococcus pneumoniae remains the most commonly identified pathogen in CAP.4,5 Increasing For editorial comment see page 1153 penicillin resistance in this organism has contributed to a shift away from penicillin as empiric therapy for CAP, particularly in patients ill enough to require hospitalization.6 However, if the S pneumoniae isolated is fully sensitive, penicillin is highly effective and cheaper than other alternatives, such as quinoClinical Investigations

lones or third-generation cephalosporins. Even with a minimum inhibitory concentration (MIC) of 1.0 mg/mL, penicillin is still likely to be effective.7 Unless physicians are prepared to change to less expensive antibiotic therapy when appropriate based on blood culture results, a cost savings will not occur. As our community has a high prevalence of penicillin-resistant S pneumoniae (PRSP), we reviewed the impact of blood cultures positive for pneumococcus on CAP management in our hospital. Materials and Methods The medical records of all patients admitted to Methodist Healthcare Central Hospital in Memphis between January 1996 and December 1998 with an admission diagnosis of CAP and at least one positive blood culture for S pneumoniae were retrospectively reviewed. In addition to demographic data, the initial antibiotic treatment, culture results, subsequent modification of treatment, complications, and outcome were recorded. Clinical data sufficient to calculate both the APACHE (acute physiology and chronic health evaluation) II score8 and the Pneumonia Severity Index (PSI) score as defined by Fine et al9 were also obtained. A change in management was defined as any change in the dose of antibiotic, or the addition or discontinuation of one or more antibiotics. If the treating physician documented in the patient’s chart that the change in management was in response to the culture result, then the change was regarded as definitely related. When a change in management occurred without any clear documentation being made in the chart to explain the reasons for the change, then the case was reviewed independently by two pulmonary physicians (GW and RW). As there was no disagreement between the two reviewers, no further review process was required. Effective coverage of atypical organisms was defined as an antibiotic regimen that would cover Legionella sp, Mycoplasma pneumoniae, and Chlamydia pneumonia. Microbiological data regarding all pneumococcal isolates during the period of interest was obtained from computerized laboratory records. Isolates were classified as penicillin-nonsusceptible S pneumoniae (PNSP) if the minimum inhibitory concentration (MIC) was $ 0.1 mg/mL; isolates were classified as PRSP if the MIC was $ 2.0 mg/mL. The total number of CAP admissions to our hospital during the review period was obtained from computerized discharge coding records. APACHE II and PSI scores are expressed as mean (SD). Differences in APACHE II and PSI values between identified groups were calculated using a two-tailed Student’s t test, with p # 0.05 considered significant.

Results During the review period, 1,805 patients were admitted with a diagnosis of CAP. Of 118 patients identified with positive blood culture results for S pneumoniae, 105 charts were available for review. Seventy-four of these patients (42 female and 32 male) were admitted for CAP; these comprised the study group. Nine of these patients (12.1%) reported

a penicillin allergy. Sources of the pneumococcal septicemia in the other 31 patients were meningitis,8 infective endocarditis (2 patients), other (8 patients), and unknown (13 patients). The choice of initial empiric antibiotic therapy is shown in Table 1. Fifty patients (67.6%) received a third-generation cephalosporin, either alone or in combination therapy, while 38 patients (51.4%) received a macrolide. No patient received empiric penicillin therapy. Fifty-eight patients (78.4%) received antibiotic therapy sufficient to cover atypical organisms. Isolates from 15 patients (11 female and 4 male; 20.3%) demonstrated either PNSP (11 patients) or PRSP (4 patients). The MIC values for penicillin of these four PRSP isolates were 2 mg/mL, 3 mg/mL, 3 mg/mL, and 4 mg/mL, respectively. Two of these patients had empiric therapy to which the isolate was resistant (cefotaxime and azithromycin). The other two patients received empiric therapy with a quinolone (levofloxacin and ofloxacin). Of the four PRSP isolates, two were reported as intermediate-grade cephalosporin resistance (MIC, 1 mg/mL) and two were classified as highly resistant to cephalosporins (MIC, 4 mg/mL). Only 1 patient (6.7%) of the 15 patients with some degree of penicillin resistance died, compared to five deaths in the 59 patients (8.5%) with penicillin sensitive isolates. This patient had a PNSP isolate that was fully sensitive to cephalosporins, and she received empiric therapy with cefotaxime. Seven isolates (9.5%) were classified as resistant to erythromycin (MIC, $ 4 mg/mL); all of these patients survived. The treating physicians did not order acute or convalescent serology for atypical pathogens on any patient. All deaths occurred in patients with PSI scores of grade V (four patients) or grade IV (1 patient). The PSI scores of all patients, with and without PNSP/ PRSP, are shown in Table 2. Mean APACHE II

Table 1–Empiric Antibiotic Therapy Therapy

Patients, No.

Cephalosporin only Cephalosporin 1 macrolide Cephalosporin 1 macrolide 1 other Cephalosporin 1 quinolone Cephalosporin 1 quinolone 1 other Cephalosporin 1 vancomycin Cephalosporin 1 other Ampicillin/sulbactam 1 macrolide Macrolide only Macrolide 1 other Quinolone only Quinolone 1 other

7 22 6 3 3 5 4 3 5 2 11 3

CHEST / 116 / 5 / NOVEMBER, 1999

1279

Table 2–PSI Scores by Penicillin Resistance Status PSI Grade Resistance

I

II

III

IV

V

Total

PNSP/PRSP Non-PNSP/PRSP Total

2 7 9

5 9 14

1 10 11

7 16 23

0 17 17

15 59 74

scores were not significantly different between patients with isolates fully sensitive to penicillin and those with any degree of penicillin resistance (14.0 [6.4] vs 13.8 [4.8]; p 5 0.83). There also was no difference between the two groups for the mean total number of points on the PSI (106.9 [39.9] vs 88.3 [34.2]; p 5 0.13). Positive pneumococcal blood culture results altered management in 31 patients (41.9%); however, antibiotic resistance was the reason for the change in management for only 2 patients. The treating physician documented that the change was due to the blood culture result in 23 patients. In the other eight cases without clear documentation in the patient chart, both independent reviewers considered the change in antibiotic therapy was mostly likely due to a culture result. In four additional cases, the treating physician documented a reason other than a blood culture result for a change in management. In two cases, an antibiotic was added for a presumed urinary tract infection. In the other two cases, azithromycin was discontinued on the fifth day of hospital admission, 48 h after the culture result became available, due to the completion of a predetermined 5-day course of therapy. In both of these cases, azithromycin had been given in addition to cephalosporin, and the cephalosporin was continued after the azithromycin was stopped. Alterations in management are summarized in Table 3. More than one change occurred in some patients. Fifty-one patients without a penicillin allergy grew penicillin-sensitive pneumococci, but antibiotic therapy was changed to penicillin in only 11 of these

Table 3–Changes in Management Following Blood Cultures* Change

Patients, No.

Macrolide stopped Change to penicillin Other stopped† Add vancomycin Change dose of cephalosporin

13 11 8 4 3

*Total is . 31 because some patients had more than one change in management. †Other 5 noncephalosporin, nonquinolone, nonmacrolide. 1280

patients (21.6%). Of the 35 patients who received a second antibiotic specifically to cover atypical organisms in addition to an antibiotic covering pneumococcus, 13 patients (37.1%) had combination therapy discontinued. Discussion Not only does the frequent indiscriminate use of broad-spectrum antibiotics increase the cost of treating CAP, it is also very likely a major factor in the growing problem of antibiotic resistance in many bacterial pathogens, particularly Gram-negative organisms such as Pseudomonas aeruginosa and Acinetobacter sp.10 Penicillin, when the pneumococcal strain identified is fully sensitive, is at least as efficacious as the higher-cost broader spectrum agents, such as third generation cephalosporins or quinolones, more commonly used as empiric therapy. The usual reason for broad-spectrum therapy of CAP is the difficulty in determining the microbiological cause. When positive, blood culture results can certainly provide useful information on both the causative agent and antibiotic sensitivity. This is particularly true for patients with less common pathogens such as Gram-negative bacteria. This study shows that in our institution, physicians were reluctant to narrow antibiotic therapy for CAP, even when blood culture results indicated this was appropriate. Therapy was switched to penicillin in only 21.6% of eligible cases, while atypical organism coverage was dropped in only 37.1%. This finding significantly reduced the impact of blood cultures on management. Several potential explanations for this reluctance exist. Multiple pathogens may be present in up to 50% patients with CAP,11–13 with M pneumoniae with S pneumoniae a frequent combination. With increasing recognition of coinfection, physicians may be reluctant to stop coverage of atypical pathogens, despite the uncertain importance of antibiotic treatment of M pneumoniae.13 As none of the patients in our review had acute or convalescent serology, we were not able to ascertain the frequency of coinfection. This finding was not surprising, since with the escalating cost of health care it is difficult to justify the cost of performing serology to patients when it rarely impacts on management. Physicians may also have been reluctant to reduce therapy because of medical/legal concerns. As this was a retrospective review, in some cases the results of blood cultures were possibly not noticed or were not noticed until after discharge. At least two patients in our review were discharged home prior to the date recorded on the blood culture report. Clinical Investigations

Although this review was limited to one large private hospital, we would be surprised if these problems are not more widespread. A recent review of patients enrolled in the multicenter Pnemonia Patient Outcomes Research Team study conducted in the United States and Canada14 found wide variations in the choice of initial antibiotic and cost of antibiotic therapy, without any apparent benefit from the higher cost regimens. Bacteremia has been identified as a risk factor for poor outcome in several studies, as reflected in the meta-analysis by Fine and colleagues.15 We expected the large majority of our study population to have PSI scores in the grade IV or V category. Surprisingly, nearly half of our patients (45.6%) were in grades I, II, or III. The absence of any significant difference in PSI or APACHE II scores between patients with sensitive or resistant isolates (PRSP or PNSP) suggests that pneumonia severity should not be considered to be a predictor of penicillin resistance. As pneumococcal penicillin resistance is more commonly identified in pediatric populations,16 the greater prevalence of PNSP/PRSP in our female patients may have been due to acquiring PNSP from contact with children or grandchildren. This female predominance has not been present in other studies.17,18 In order to reduce the cost of treating patients with CAP and to potentially alleviate the increasing problem of antibiotic resistance, physician fears about narrowing antibiotic therapy and altering prescribing practices must be addressed. This will likely require further investigation of the frequency and significance of multipathogen CAP and the education of both physicians and patients regarding the benefits of simplifying antibiotic therapy. While there is no doubt that a positive blood culture result can provide very useful information, in an era of escalating costs and rationalization of health care there is an urgent need for prospective studies on their utility and cost-effectiveness in the management of CAP. References 1 Chalasani NP, Valdecanas MA, Gopal AK, et al. Clinical utility of blood cultures in adult patients with communityacquired pneumonia without defined underlying risks. Chest 1995; 108:932–936

2 Woodhead MA, Arrowsmith J, Chamberlain-Webber R, et al. The value of routine microbial investigation in communityacquired pneumonia. Respir Med 1991; 85:313–317 3 Hickey RW, Bowman MJ, Smith GA. Utility of blood cultures in pediatric patients found to have pneumonia in the emergency department. Ann Emerg Med 1996; 27:721–725 4 Mundy LM, Auwaerter PG, Oldach D, et al. Communityacquired pneumonia: impact of immune status. Am J Respir Crit Care Med 1995; 152:1309 –1315 5 Steinhoff D, Lode H, Rudeschel G, et al. Chlamydia pneumoniae as a cause of community-acquired pneumonia in hospitalized patients in Berlin. Clin Infect Dis 1996; 22:958 – 964 6 Neiderman MS, Bass JB, Campbell GD, et al. Guidelines for the initial management of adults with community acquired pneumonia: diagnosis, assessment of severity, and initial antimicrobial therapy. Am Rev Respir Dis 1993; 148:1418 – 1426 7 Pallares R, Gudiol F, Linares J, et al. Risk factors and response to antibiotic therapy in adults with bacteremic pneumonia caused by penicillin-resistant pneumococci. N Engl J Med 1987; 317:18 –22 8 Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13:818 – 829 9 Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997; 336:243–250 10 Gold HS, Moellering RC. Antimicrobial-drug resistance. N Engl J Med 1996; 335:1445–1453 11 Lieberman D, Schlaeffer F, Boldur I, et al. Multiple pathogens in adult patients admitted with community-acquired pneumonia: a one year prospective study of 346 consecutive patients. Thorax 1996; 51:179 –184 12 Mundy LM, Oldach D, Auwaerter PG, et al. Implications for macrolide treatment in community-acquired pneumonia. Chest 1998; 113:1201–1206 13 Lieberman D, Schlaeffer F, Lieberman D, et al. Mycoplasma pneumoniae community-acquired pneumonia: a review of 101 hospitalized patients. Respiration 1996; 63:261–266 14 Gilbert K, Gleason PP, Singer DE, et al. Variations in antimicrobial use and cost in more than 2000 patients with community-acquired pneumonia. Am J Med 1998; 104:17–27 15 Fine MJ, Smith MA, Carson CA, et al. Prognosis and outcomes of patients with community-acquired pneumonia: a meta analysis. JAMA 1996; 275:134 –141 16 Labowitz A, Young A, Heffernan R, et al. Surveillance for penicillin-nonsusceptible Streptococcus pneumoniae: New York City 1995 [letter]. JAMA 1997; 277:1585 17 Moreno F, Crisp C, Jorgensen JH, et al. The clinical and molecular epidemiology of bacteremias at a University hospital caused by pneumococci not susceptible to penicillin. J Infect Dis 1995; 172:427– 432 18 Hofmann J, Cetron MS, Farley MM, et al. The prevalence of drug-resistant Streptococcus pneumoniae in Atlanta. N Engl J Med 1995; 333:481– 486

CHEST / 116 / 5 / NOVEMBER, 1999

1281