| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Laboratory Medicine, National Defense Medical College, Tokorozawa, Saitama, Japan
2 Department of Internal Medicine, Nishi-Ohmiya Hospital, Saitama-City, Saitama, Japan
3 Department of Information Technology and Decision Sciences, Yamaguchi University School of Medicine, Ube, Yamaguchi, Japan
aaddress correspondence to this author at: Department of Laboratory Medicine, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan. Fax 81-42-995-0633, e-mail yutakemu{at}interlink.or.jp
Physicians often prescribe antibiotics to febrile patients despite a lack of evidence of bacterial infection (1)(2)(3)(4)(5). C-Reactive protein (CRP) and leukocyte count [white blood cell count (WBC)] can contribute to differentiation of possible bacterial vs viral infections (6) because patients with acute bacterial infections typically have markedly increased CRP and WBC whereas patients with viral infection do not. Other studies (7)(8)(9) have also suggested the potential value of CRP with WBC in patients with infections or inflammatory conditions. We studied the potential usefulness of immediate testing for CRP and WBC in physicians’ decisions to prescribe antimicrobial drugs for acutely febrile new outpatients.
This study was carried out in the mornings in a primary care setting. Among all new outpatients who visited the General/Internal Medicine Clinic of Nishi-Ohmiya Hospital (a regional/community 150-bed hospital treating 
500 outpatients per day), 305 patients presenting with fever (
37.5 °C) <8 days in duration and suspected of having infection were entered in this study. Eleven physicians, ages 29–53 years with 5–29 years of experience, participated in patient clinical examination during the study period from December 2000 to January 2003. Three physicians were full-time members of the hospital staff. Patients were randomized by a study controller into two groups irrespective of symptoms. As illustrated in Fig. 1 in Data Supplement 1, which accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol50/issue1/, one group (147 patients) underwent CRP and WBC testing before the physician’s initial clinical examination (advance testing group), and physicians made the initial clinical diagnosis and management decisions after test results were reported. The other group (154 patients; without advance testing group) did not receive tests before the initial clinical examination, and diagnoses and management decisions were based on history and physical examination unless urgent tests were ordered. The numbers of patients with and without advance testing for each physician were nearly equal, except for a few physicians who saw few patients. The study design did not restrict in either patient group urgent testing considered necessary by the physician. The results of nonurgent tests ordered at the initial clinical examination (additional or subsequent tests) were evaluated at the patient’s next visit. All patients were informed of the study design and registered only after providing written informed consent to the study. The protocol was approved by the hospital’s Clinical Study Committee.
CRP was measured by nephelometry (Model TBA-30FR; Toshiba) and WBC by an automated blood cell counter (Model KX-21; Sysmex). Approximately 40–50 min were required to separate serum and obtain a CRP result, and 10 min were needed for a WBC measurement, provided that the analyzers were ready to use. Between-run imprecision (CV) was 3.3–6.3% at 5–22 mg/L for CRP and 5.3–5.9% at 2.5–6.8 x 109/L for WBC. The reference intervals used by physicians were 
5 mg/L for CRP and 3.5–9.0 x 109/L for WBC in the diagnosis and differentiation of infection.
We built a patient database comprising the results of advance testing, initial clinical diagnosis based on all available information at the initial clinical examination, differentiation of infection by physicians, prescription menu, physicians’ decisions for patient management, and test items subsequently ordered. Data were analyzed with statistical comparisons by the 
2 test or the Fisher exact test for categorical and binary variables. Distributions of continuous data were analyzed by the Mann–Whitney U-test or the Student t-test (StatFlex 5.0; Artec Inc.). All statistical tests were two-tailed.
Of 151 patients randomized into the group receiving advance testing of CRP and WBC, 4 refused to receive testing, thus giving 147 patients with advance testing and 154 without advance testing enrolled in the present study. Table 1
shows the demographic and clinical characteristics of the 301 study patients. In total, there were 227 and 241 visits by patients with and without advance testing, respectively. The reconsultation rate was slightly higher in patients with advance testing than in those without advance testing. However, early revisits (within 3 days after the first clinical examination), which suggested no improvement in the patient’s physical condition and persistent infection despite starting treatment, were similar for both groups: 21 patients (48%) with advance testing vs 20 patients (56%) without advance testing. Other patient clinical characteristics were almost identical between the two groups. In patients with advance testing, CRP ranged from 0 to 226 mg/L (median, 19 mg/L), and WBC ranged from 1.6 to 21.4 x 109/L (median, 6.7 x 109/L).
|
There was a consensus in the differentiation of infection among participating physicians: a condition with highly increased CRP concomitant with leukocytosis (e.g., WBC count 9 x 109/L) was likely attributable to a bacterial pathogen, whereas slight or moderate increases in CRP with normal or decreased WBC values suggested a viral infection. Physicians’ estimation of infection was valid when the treatment that followed was concordant with the presumptive infection. Availability of CRP and WBC data at the first patient visit led to a differentiation of infection in 97 of 147 patients (66%), comprising 39 bacterial and 58 nonbacterial infections. Conversely, history and physical findings alone could differentiate infections in only 19 patients (12 bacterial and 7 viral infections) in the without-advance-testing group (see online Data Supplement 2). Table 1
also shows the rates at which physicians prescribed antimicrobial drug in the two patient groups. Notably, 89 patients (61%) with advance testing received antimicrobial drugs, whereas 141 (92%) without advance testing were administered various kinds of antimicrobial drugs. In the former patient group, 76 patients (52%) received antibiotics: 62 received oral antibiotics, 3 received intravenous antibiotic infusion, and 11 received both. Thirteen patients were prescribed an oral antiviral agent (such as amantadine, zanamivir, and oseltamivir) based on the diagnosis of probable influenza infection. In 14 patients who received intravenous antibiotic infusion in the advance testing group, 13 showed a definite leukocytosis (WBC count >9 x 109/L) and/or a marked increase in CRP (>100 mg/L) in addition to a severe physical condition. In the without-advance-testing group, however, antibiotics were administrated to 135 patients (88%): 122 received oral antibiotics and 13 received both oral and parenteral antibiotics. Therefore, a reduction of nearly 40% in prescription of antibiotics was achieved by CRP and WBC advance testing in acutely febrile patients.
We analyzed the influence of CRP and WBC values on physicians’ decisions for antimicrobial drug use in patients with advance testing. In patients with CRP 40 mg/L or WBC 9 x 109/L, 83% and 93%, respectively, received antibiotics, whereas only 37% and 35% of patients who had CRP <40 mg/L or WBC count <9 x 109/L were given antibiotics. On the other hand, 10 of 13 patients who were administered an antiviral agent for probable influenza infection had CRP <40 mg/L, and 11 of them had a WBC count <9 x 109/L. Only one in four patients who had CRP <40 mg/L and WBC count <9 x 109/L received antibiotic treatment, whereas 65% of such patients were not given any antimicrobial drugs (see online Data Supplement 3).
In this study, we analyzed physicians’ decisions to prescribe antimicrobial drugs for patients suspected of having an acute infection when notified of CRP and WBC results at the initial clinical examination. Immediate availability of CRP and WBC test data led to a successful reduction in the prescribing of antibiotics, compared with the patient group without advance testing. In contrast, administration of antiviral agents for probable influenza infection increased in the advance testing group. Of 147 patients who received CRP and WBC advance testing, infection could be differentiated to possible bacterial or nonbacterial origin after test data were interpreted for 97 patients, of whom 52 received the appropriate antimicrobial treatment against the corresponding infection estimated, whereas 45 were not treated with any antimicrobial drug because the infections were estimated to be self-limiting and nonbacterial. Although differentiation of infections based solely on CRP and WBC data may not necessarily lead to a microbiologically accurate diagnosis and treatment, physicians treated 66% of patients based on CRP and WBC values with or without prescribing antimicrobial drugs. On the other hand, physical findings alone allowed differentiation of only 12 patients with probable bacterial infection from the entire without-advance-testing patient group, although a majority of patients (135 of 153) were prescribed antibiotics (Table 1
).
In patients who had multiple consultations, rates for early revisits were similar in the two patient groups. In 5 of the 21 patients in the advance testing group who revisited the clinic within 3 days, physicians decided to commence oral or intravenous antibiotic administration at the second consultation because of their unimproved physical conditions, whereas they did so in 6 of the 20 without-advance-testing patients who had early reconsultations. The lack of an increment in the rate of the apparent initial treatment failures in the advance testing patient group allows us to conclude that antibiotic reduction in this group did not adversely affect the treatment outcome.
There is indeed a diagnostic dilemma in differentiating bacterial and nonbacterial infections for proper use of antimicrobial drugs. Standard diagnostic procedures take too long for an immediate decision (10). Primary care physicians therefore commonly prescribe antibiotics to many of their patients suspected of having acute infections, but without definite evidence. In such patients, it is difficult to distinguish bacterial infections based on physical findings alone, as demonstrated in our patients without advance testing. In addition, patients presenting with symptoms of infection often believe that bacterial infection is the problem and that antibiotics are the answer (11). Such patient expectations are likely to induce physicians to prescribe antibiotics. Indeed, it has been reported that general practitioners prescribed antibiotics in three-fourths of such patients, labeling many as bacterial infection (1)(2). Increasing antibiotic use contributes to rising drug costs and increases antibiotic resistance of causative microorganisms in the community (12)(13). We demonstrated here that testing for only two inflammation markers was able to change physicians’ decisions concerning prescription of antimicrobial drugs. In addition, test data are likely to play a powerful persuasive role for patients expecting to be prescribed antibiotics.
This study was designed in the manner of near-patient testing (the hospital clinical laboratory was located adjacent to the outpatient unit where patient evaluations were carried out), although we did not use a discrete instrument for CRP analysis. In most patients, CRP and WBC results were available within 1 h after the time of blood drawing. Our results indicate the importance of immediate testing of CRP and WBC for proper use of antimicrobial drugs. Some investigators have emphasized the potential usefulness of near-patient measurement of CRP for proper treatment of patients with infections in primary care practice (10)(14)(15). Although testing costs would be compensated by the reduction in drug costs, the advance testing approach may not necessarily translate into savings in total cost. In fact, the decrease in costs for antibiotic prescriptions in the advance testing group gave savings of ¥56 400 (US$513) compared with the without advance testing group, whereas CRP and WBC testing required ¥83 400 (US$758) if the testing costs were calculated based cost data in our previous study (8). Advance testing decreased the additional testing that followed (see online Data Supplement 2); however, frequent use of newly developed drugs such as antiviral agents, which are currently more expensive than the older generation of antibiotics, may lead to an increment in drug costs. Nevertheless, such financial considerations would not negate the essential benefit produced by the immediate testing, which contributes to physicians’ timely and proper decisions for patient management. Because testing costs would likely differ between a strict point-of-care testing model and our study, in which automated multichannel analyzers were used, the financial aspects of advance CRP and WBC testing will be carefully analyzed in a future study, which will be carried out with strict adherence to point-of-care testing protocol.
In conclusion, immediate availability of CRP and WBC test results at the first clinical examination led to a dramatic reduction in antibiotic use. Furthermore, onsite CRP and WBC testing is more likely to contribute to appropriate use of antimicrobial drugs in patients with acute infections. This pilot study should be extended in the future, with larger numbers of patients and participating physicians in multiple primary care settings.
Acknowledgments
This study was supported in part by grants from the International Clinical Pathology Center (Tokyo, Japan). We are grateful to Mina Sasaki for her work as the study controller, and to Akiko Nishioka, Yukako Wachi, Fumie Katano, and Misako Funatsu for immediate testing of CRP and WBC. We thank Kuniko Okada for collection of patient blood, Satomi Madarame for data assembly, and Hiroki Saitoh for drug cost calculations. We also thank Drs. Nobukazu Sasaki, Hironari Kimura, Yoshinari Moriyama, Mika Tsuiki, and Hiroki Kawahara for their participation in the clinical practice.
Editor’s Note: We note that the practice of nearly routine prescribing of antibiotics for fever is not universal.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  Y Takemura, K Ebisawa, H Kakoi, H Saitoh, H Kure, H Ishida, and M Kure Antibiotic selection patterns in acutely febrile new outpatients with or without immediate testing for C reactive protein and leucocyte count J. Clin. Pathol., July 1, 2005; 58(7): 729 - 733. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
