Clinicalpharmacy.ucsf.edu

Doxycycline Is a Cost-effective Therapy
for Hospitalized Patients With
Community-Acquired Pneumonia
Reba K. Ailani, MD; Gautami Agastya, MD; Rajesh K. Ailani, MD; Beejadi N. Mukunda, MD; Raja Shekar, MD Background: Doxycycline has a high degree of activ-
tibiotic and the clinical response was 2.21 ± 2.61 days in ity against many common respiratory pathogens and has the doxycycline group compared with 3.84 ± 6.39 days been used in the outpatient management of lower respi- in the control group (P = .001). The mean ± SD length ratory tract infections, including pneumonia.
of hospitalization was 4.14 ± 3.08 days in the doxycy-cline group compared with 6.14 ± 6.65 days in the con- Objective: To evaluate the efficacy of intravenous doxy-
trol group (P = .04). The median cost of hospitalization cycline as empirical treatment in hospitalized patients with was $5126 in the doxycycline group compared with $6528 mild to moderately severe community-acquired pneumonia.
in the control group (P = .04). The median cost of anti-biotic therapy in the doxycycline-treated patients ($33) Patients and Methods: We conducted a randomized
was significantly lower than in the control group prospective trial to compare the efficacy of intravenous ($170.90) (PϽ.001). Doxycycline was as efficacious as doxycycline with other routinely used antibiotic regi- the other regimens chosen for the treatment of commu- mens in 87 patients admitted with the diagnosis of com- munity-acquired pneumonia. Forty-three patients wererandomized to receive 100 mg of doxycycline intrave- Conclusion: Doxycycline is an effective and inexpen-
nously every 12 hours while 44 patients received other sive therapy for the empirical treatment of hospitalized antibiotic(s) (control group). The 2 patient groups were patients with mild to moderately severe community- comparable in their clinical and laboratory profiles.
Results: The mean ± SD interval between starting an an-
PNEUMONIAISthesixthlead- Invariousstudiesfromtheliterature,
percentage of cases where no definitive cause for CAP was established ranged from 32.9%3 to 63.3%.4 The most commonly detected or- ganism was Streptococcus pneumoniae, rang- treating the illness in United States is es- monias caused by Legionella, Chlamydia, and Mycoplasma together accounted for 8% to likely causative organisms until a micro- pathogens including S pneumoniae, Hae- biologic cause is established. However, a mophilus influenzae, Moraxella catarrha- lis, anaerobes such as Bacteroides and an- atypical agents like Legionella, Myco- plasma pneumoniae, and Chlamydia pneu- moniae.7 Doxycycline has been used as oral lower respiratory tract infections (includ- ics should include the following: antibac- terial spectrum, efficacy, adverse effects, has been found to be comparable.8-10 Doxy- cycline is an attractive alternative for hos- 1999 American Medical Association. All rights reserved.
PATIENTS, MATERIALS,
cline group or the control group depending on the num- AND METHODS
The Institutional Research and Review Committee of the hospital approved the study protocol. All patients were The study was conducted in a 371-bed community teach- informed about the study and a written consent was ob- ing hospital from August 1995 to December 1997. Patients were eligible for the study if they had a clinical and radio-logical diagnosis of pneumonia acquired in the community PATIENTS
before admission to the hospital. Patients were excluded ifany of the following were present: younger than 18 years, After enrollment, patients were interviewed and informa- pregnant or lactating women, history of allergic reaction to tion was recorded on a standard data collection form. Data the use of tetracycline or doxycycline, severe hepatic or re- included demographic information, history, physical ex- nal dysfunction, human immunodeficiency virus infection, amination, and admission laboratory findings. All pa- immunocompromised state, clinical sepsis, patients requir- tients had baseline investigations that included a chest ra- ing intubation, and patients from a nursing home or a long- diograph, complete blood cell count with differential, serum term care facility. Patients were enrolled in the emergency electrolytes, serum urea nitrogen, serum glucose, pulse ox- department or soon after admission to the hospital.
imetry or arterial blood gases, sputum Gram stain and cul- During the study period there were a total of 356 pa- ture, and blood cultures. All patients were examined daily tients admitted with the diagnosis of pneumonia. Of these for their complete blood cell counts and electrolyte levels.
356 patients, 136 were from long-term care facilities. Forty- Additional tests such as sputum, acid-fast stains, serology six patients required intubation in the emergency depart- for Chlamydia, Mycoplasma, and Legionella were per- ment. Thirty-eight had 1 of the exclusion criteria includ- formed when clinically indicated, at the discretion of the ing immunocompromised state, renal or hepatic failure, or pregnancy. Another 22 patients received antibiotics be-fore enrollment. In 27 patients who were not enrolled in COMORBID CONDITIONS
the study, consent could not be obtained. Eighty-seven pa-tients hospitalized with CAP were randomly assigned to re- Conditions such as chronic obstructive pulmonary dis- ceive either doxycycline or 1 or more antibiotics selected ease, coronary artery disease, congestive heart failure, neo- plastic disease, diabetes mellitus, chronic renal failure, al- The randomization process was done by a predeter- coholism, and smoking were recorded.
mined numbered sequence. As the patients were enrolledinto the study, they were assigned to either the doxycy- pitalized patients with mild to moderately severe CAP.
cycline group, compared with 3.84 ± 6.39 days in the The objective of this study was to evaluate the efficacy control group (P = .001) (Table 3). The mean ± SD of intravenous doxycycline as empirical treatment in hos- length of hospitalization was 4.14 ± 3.08 days in the pitalized patients with mild to moderately severe CAP.
doxycycline group, compared with 6.14 ± 6.65 days inthe control group (P = .04) (Table 3). The mean ± SD number of antibiotics used, including the one given ondischarge, was significantly lower in the doxycycline A total of 87 patients were enrolled into the study. Forty- group than in the control group: 1.16 ± 1.04 in the three were randomized to receive doxycycline, and 44 doxycycline group vs 2.43 ± 1.59 in the control group were assigned to the control group. The clinical and labo- ratory features between the 2 groups of patients were not In the doxycycline group, 6 patients developed ad- significantly different, as shown in Table 1 and Table 2.
verse effects, but none required a change of antibiotic be- The ages of the patients in the doxycycline and con- cause of the adverse effects. In the control group, 11 pa- trol groups ranged from 18 to 84 years, and 23 to 81 years, tients had adverse effects (P = .19) and 3 of these 11 respectively (Table 1). However, only 20 (23%) of the required a change of antibiotic. Two patients had diar- 87 patients were older than 60 years. Twenty-seven pa- rhea with the use of clindamycin and 1 patient had se- tients (63%) in the doxycycline group and 21 (48%) in vere vomiting with the use of erythromycin; hence, their the control group were women. The average number of comorbid conditions between the 2 groups were not sta- Three patients in the doxycycline group required a tistically significant (P = .18). In both groups, 40% of the change in antibiotic because of a lack of response. Two patients had documented underlying lung disease.
patients failed to respond to the use of doxycycline. One The median cost of antibiotic(s) during hospital- patient was misdiagnosed and had tuberculosis on the ization for the doxycycline group was $33, compared basis of a strongly positive tuberculin skin test and a per- with $170.9 in the control group (PϽ.001) (Table 3).
sistent infiltrate on chest radiograph. This was not con- The median cost of hospitalization in the doxycycline firmed by cultures; however, the patient improved with group was $5126, compared with $6528 in the control group (P = .04) (Table 3). The mean ± SD time to Eight patients in the control group required a respond to treatment was 2.21 ± 2.61 days in the doxy- change in the antibiotic regimen, 3 because of adverse 1999 American Medical Association. All rights reserved.
was 100 mg of doxycycline orally every 12 hours. The at- TREATMENT
tending physician determined the time of discharge.
Doxycycline (100 mg) was given intravenously every 12 STUDY CRITERIA
hours in patients randomized to the study group. Inpatients randomized to the control group, the physician The 2 patient groups were then compared on the following taking care of the patient chose antibiotics without parameters: (1) time to resolution of morbidity, (2) length restrictions. All patients were monitored daily by one of of hospital stay, (3) cost of antibiotics during hospitaliza- us for improvement or deterioration while receiving tion, (4) cost of hospitalization, (5) adverse effects from the therapy and were also monitored for any adverse drug use of antibiotics, and (6) number of antibiotics used per pa- tient, including the antibiotic on discharge. The time to re-spond to treatment was calculated as described earlier. The CRITERIA FOR RESPONSE
length of hospital stay was calculated by subtracting the ad-mission date from the discharge date. The cost of antibiot- The time to respond to treatment was defined as the ics used during hospitalization was obtained from the num- number of days between the day of randomization (day ber of antibiotic doses times the cost of each dose. The cost 1) and the day on which the last of the following param- of each dose of antibiotic was obtained from the minifor- eters was met: (1) oral temperature of 37.9°C or lower mulary of the pharmacy department of the hospital and was over 3 consecutive 8-hour periods; (2) beginning of a the amount charged to the patient. The mean cost of anti- decrease in total white blood cell counts, ie, the day the biotic in the doxycycline and the control groups included count showed a tendency toward normal in patients who the antibiotics that were used even when the patient failed had leukocytosis; (3) subjective improvement of symp- while receiving initial therapy. The cost of hospitalization toms for which the patient was admitted to the hospital.
was obtained from the computer records of the hospital bill This was decided by the primary physician’s clinical notes and interview of the patient by one of us; and (4)in patients with no underlying disease like chronic STATISTICAL ANALYSIS
obstructive pulmonary disease, congestive heart failure,and who were hypoxic on admission, resolution of Comparison of patient age was done by Student 2-sample t test. All other variables were analyzed by Wilcoxon rank Patients who showed any sign of deterioration or sum test rather than the Student t test, because of viola- who were not responding to therapy or who developed se- tion of the distributional normality assumption. Because rious adverse effects had their antibiotic regimen changed of nonnormal distributions exhibited by most of these vari- at the discretion of the physician taking care of them. Pa- ables, medians along with the means and SDs were calcu- tients who improved while receiving therapy were lated. The adverse effects were compared by the ␹2 test.
switched to an oral regimen, which in the study group PϽ.05 was considered to indicate statistical significance.
effect (as described earlier) and 5 because of a lack of moderately severe CAP. Doxycycline was more cost- effective than other antibiotic options chosen for the em- Of the 87 patients enrolled in the study, 7 were found pirical treatment of CAP. The cost of hospitalization, time to have pneumococcal bacteremia. Three of these 7 pa- to respond to treatment, and the length of hospitaliza- tients received doxycycline and 4 received other antibi- tion were significantly lower in the doxycycline group otics. The pneumococci were sensitive to the use of peni- than in the control group. Besides being inexpensive and cillin in 6 of the 7 patients. One patient had intermediate- effective, the number of patients with adverse effects in level resistance to penicillin and he had been randomized the doxycycline group was less than in the control group.
to the control group. All these isolates were susceptible This was probably because many patients in the control to the use of tetracycline. All 7 patients recovered and group received complex regimens involving 2 or more From January to June 1997 in our hospital system, The empirical treatment of CAP has many options.
10% of S pneumoniae isolates had high-level resistance Some of the options involve the use of 2 agents to cover to penicillin (minimal inhibitory concentration, Ն2 µg/ the common respiratory pathogens like S pneumoniae, H mL) and another 10% had intermediate-level resistance influenzae, as well as organisms that cause atypical pneu- to penicillin (minimal inhibitory concentration, 0.1-1.0 monia. According to the recommendations of the Infec- µg/mL). Resistance to tetracycline was 6.5%. Of the peni- tious Diseases Society of America,12 the empirical treat- cillin-resistant pneumococci, 16.6% were resistant to tet- ment for patients hospitalized to the general wards racycline (B.N.M., written communication, May 18, 1998).
includes a ␤-lactam antibiotic with or without a macro-lide or a quinolone.12 Doxycycline has been recom- mended only in the empirical treatment of outpatientswith CAP.12 Even in the recommendations of the Ameri- Doxycycline has a good activity against most of the com- can Thoracic Society,13 tetracycline has been recom- mon pathogens causing CAP.7,11 In this study, we have mended only as an outpatient treatment of CAP.
shown that doxycycline is an effective option for the treat- Historically, tetracycline is considered inferior to peni- ment of patients admitted to the hospital with mild to cillin in the treatment of S pneumoniae infections. How- 1999 American Medical Association. All rights reserved.
Table 1. Clinical Features Comparing the
Table 3. Response to Treatment: Comparison Between
Doxycycline-Treated Patients and Control Patients
the Doxycycline-Treated Patients and Control Patients*
Doxycycline
Doxycycline Group
Control Group
Clinical Features
Table 2. Laboratory Features Comparing the Doxycycline
and Control Groups*
Doxycycline
Laboratory Features
S pneumoniae from Franklin County, Ohio, between 1991 and 1994. Doxycycline was found to have excellent activ- ity against pneumococcus (99.2% isolates susceptible) with no trend toward decreasing susceptibilities. High-level peni- cillin resistance was found in 5 isolates, and all of them were susceptible to doxycycline. Among isolates with interme- diate level of sensitivity to penicillin, only 6% were not sus- Our study shows doxycycline to be effective for the empirical treatment of hospitalized patients with mild to Some limitations of this study include (1) the study was not blinded, (2) patients with severe disease requir- ing intubation were excluded, (3) the cause of pneu- monia was not established in most of our patients, and (4) we did not perform antibacterial susceptibilities of the organisms isolated against doxycycline.
Doxycycline has activity against the common res- piratory pathogens, relatively low toxicity, very low cost, and a convenient twice-daily dosing. It should be con-sidered in the empirical treatment of hospitalized pa- ever, in certain places there has been a changing trend, with tients with mild to moderately severe CAP.
decreasing tetracycline resistance and increasing penicil-lin resistance.14 In the United States, 7.5% of S pneumoniae Accepted for publication May 26, 1998. are resistant to tetracycline.15 Moreover, tetracycline resis- Presented in part as an oral research abstract at the tance is not the same as doxycycline resistance. In a study annual meeting of the American College of Physicians, Ohio of 256 clinical isolates of S pneumoniae described by Shea chapter, Columbus, August 15, 1997. and Cunha,7 30% were resistant to penicillin, 20% to tet- We thank Burton C. West, MD, for his contribution in racycline, and only 5% to doxycycline. Plouffe et al16 re- this study and help in preparing the manuscript; Adrian Cara- ported antimicrobial susceptibilities on 499 isolates of cioni, Marketa Kasalova, Ali Malick, and Jan Kasal for their 1999 American Medical Association. All rights reserved.
contribution in enrolling patients during this study; and 7. Shea KW, Cunha BA. Doxycycline activity against Streptococcus pneumoniae.
Gordon Jacobsen, statistician at the Henry Ford Health Sys- 8. Pedley JB. Treatment of acute exacerbations of chronic bronchitis in general prac- tem, Detroit, Mich, for help with the statistical analysis. tice. Br J Clin Pract. 1969; 23:280-283.
Reprints: Raja Shekar, MD, Department of Medicine, 9. Harazim H, Wimmer J, Mittermayer HP. An open randomized comparison of ofloxa- Meridia Huron Hospital, 13951 Terrace Rd, East Cleve- cin and doxycycline in lower respiratory tract infections. Drugs. 1987;34(suppl 10. Biermann C, Leken A, Riise R. Comparison of spiramycin and doxycycline in the treatment of lower respiratory infections in general practice. J Antimicrob Che- 11. Schlick W. The problems of treating atypical pneumonia. J Antimicrob Che- 1. National Center for Health Statistics. Vital Statistics of the United States, 1989: 12. Infectious Diseases Society of America. Community-acquired pneumonia in adults: Mortality, Part A. Vol 2. Washington, DC: Public Health Service; 1993. US Dept guidelines for management. Clin Infect Dis. 1998;26:811-838.
of Human and Health Services publication (PHS) 93-101.
13. American Thoracic Society. Guidelines for the initial management of adults with 2. Marie TJ. Community acquired pneumonia. Clin Infect Dis. 1994;18:501-515.
community-acquired pneumonia: diagnosis, assessment of severity, and initial 3. Fang GD, Fine M, Orloff J, et al. New and emerging etiologies for community antimicrobial therapy. Am Rev Respir Dis. 1993;148:1418-1426.
acquired pneumonia with implications for therapy: a prospective multicenter study 14. Linares J, Pallares R, Alonso T, et al. Trends in antimicrobial resistance of clini- of 359 cases. Medicine. 1990;69:307-316.
cal isolates of Streptococcus pneumoniae in Bellvitge hospital, Barcelona, Spain 4. Ostergaard L, Anderson PL. Etiology of community-acquired pneumonia: evalu- (1979-1990). Clin Infect Dis. 1992;15:99-105.
ation by transtracheal aspiration, blood culture, or serology. Chest. 1993;104: 15. Doern GV, Brueggemann A, Holly HP Jr, Rauch AM. Antimicrobial resistance of Streptococcus pneumoniae recovered from outpatients in the United States dur- 5. Marrie TJ, Durant H, Yates L. Community-acquired pneumonia requiring hospi- ing the winter months of 1994 to 1995: results of a 30-center national surveil- talization: 5-year prospective study. Rev Infect Dis. 1989;11:586-599.
lance study. Antimicrob Agents Chemother. 1996;40:1208-1213.
6. Bates JH, Campbell GD, Barron AL, et al. Microbial etiology of acute pneumonia 16. Plouffe JF, Breiman RF, Facklam RR. Bacteremia with Streptococcus pneu- in hospitalized patients. Chest. 1992;101:1005-1012.
moniae: implications for therapy and prevention. JAMA. 1996;275:194-198.
1999 American Medical Association. All rights reserved.

Source: http://clinicalpharmacy.ucsf.edu/idmp/id_lib/Community-Acquired%20Pneumonia/doxyCAP;ArchIntMed;1999;159;266.pdf