Journal of Pharmacy Practice and Pharmaceutical Sciences

ISSN: 2652-2536

Research Article

Retrospective Evaluation of Doxycycline versus Azithromycin in Non-Intensive Care Unit Veterans Hospitalized with Community-Acquired Pneumonia

Justin R Spivey1,2, Robert C Wood1,3, Heather A Sirek4and Jonathan P Moorman2,5

1Department of Pharmacy, Mountain Home VA Healthcare System, Mountain Home, Tennessee, USA

2Center of Excellence in Inflammation, Infectious Disease, and Immunity, East Tennessee State University, Johnson City, Tennessee, USA

3Bill Gatton College of Pharmacy, East Tennessee State University, Johnson City, Tennessee, USA

4Department of Pharmacy, Mayo Clinic, Rochester, Minnesota, USA

5Department of Medicine, Mountain Home VA Healthcare System, Mountain Home, Tennessee, USA

Received: 21 June 2019

Accepted: 26 July 2019

Version of Record Online:  13 August 2019


Spivey JR, Wood RC, Sirek HA, Moorman JP(2019) Retrospective Evaluation of Doxycycline versus Azithromycin in Non-Intensive Care Unit Veterans Hospitalized with Community-Acquired Pneumonia.J Pharm Pract Pharm Sci 2019(1): 66-74.

Correspondence should be addressed to
Robert C Wood, USA

DOI: 10.33513/PPPS/1901-14


Copyright © 2019 Robert C Wood. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and work is properly cited.


Purpose: To compare the efficacy of doxycycline as an alternative to azithromycin, when allergy or patient factors preclude use, each in combination with ceftriaxone for non-Intensive Care Unit (ICU) veterans hospitalized with Community-Acquired Pneumonia (CAP), which in current guidelines is recommended but is based on limited direct clinical evidence and in vitro data.

Methods: This retrospective cohort study included veterans with pneumonia admitted to the VA Midsouth Healthcare Network who received ceftriaxone plus either doxycycline or azithromycin. The primary endpoint was a composite outcome consisting of mortality or escalation of antibiotics during the index hospitalization, hospital readmission, or emergency department visit within 30 days of discharge. Univariate and multivariate logistic regression were performed to identify risk factors associated with clinical outcomes.

Results: 3,788 veterans were included with 77 in the doxycycline group and 3,711 in the azithromycin group. Participants were well-matched according to CAP severity and comorbidities. The composite outcome occurred in 40 (51.9%) and 1645 (44.3%) of veterans in the doxycycline and azithromycin groups, respectively (p=0.18). Receipt of doxycycline was not associated with components of the composite outcome in multivariate analyses.

Conclusion: Doxycycline was not associated with a statistical increase in adverse patient outcomes compared with azithromycin. This study appears to support current Infectious Diseases Society of America and American Thoracic Society guideline recommendations of doxycycline as an alternative to azithromycin as part of combination therapy for non-ICU patients hospitalized with CAP.


Azithromycin; Community-Acquired Pneumonia; Doxycycline; Veteran


Pneumonia remains a source of significant morbidity and mortality despite the availability of numerous effective in vitro antimicrobials coupled with several vaccinations aimed at prevention. The National Vital Statistics Report of deaths for 2014 described influenza and pneumonia as the eighth most common etiology of mortality in the United States. These conditions were responsible for approximately 55,000 deaths, constituting 2.1% of total deaths for this period[1]. A recently published study estimated the number of annual deaths in the US population during a Community-Acquired Pneumonia (CAP) hospitalization at 102,821 with 205,642 at 30 days post-discharge[2]. Furthermore, the incidence of pneumonia and resulting morbidity and mortality are elevated in patients with comorbidities and in those of increased age[2,3]. A 2015 national Veterans Affairs (VA) patient population analysis described a lower CAP incidence rate for persons 18 to 49 years of age compared with veterans ³50 years of age [4]. Additionally, this analysis demonstrated a linear increase in the risk of 30-day all-cause readmission and death within one year for patient age and risk group, determined by the presence of chronic medical conditions and immunosuppression[4]. Thus, optimizing therapy for this condition is paramount in patients at increased risk of adverse clinical outcomes, such as elderly patients.

The Infectious Diseases Society of America (IDSA) and American Thoracic Society (ATS) jointly released CAP guidelines in 2007 which recommend ceftriaxone in combination with doxycycline as an alternative to ceftriaxone and azithromycin for non-intensive Care Unit (ICU) patients hospitalized with CAP[5]. This guideline statement is rated as “level III (low) evidence” but represents an attractive alternative due to potential drug-drug interactions and concerns of azithromycin-related cardiovascular toxicity, particularly in patients at elevated baseline risk[6]. There is much clinical debate in the medical literature surrounding the true cardiovascular toxicity of azithromycin[6-12]. However, clinicians likely frequently make treatment decisions based on this conflicting data and utilize non-first line therapeutic regimens, such as doxycycline, in patients perceived to be at an elevated risk for toxicity.

Furthermore, the recommendation in the IDSA/ATS guidelines for doxycycline in the treatment of non-ICU in patients with CAP was made, “on the basis of scant data for treatment of Legionella infections”[5]. This recommendation was recently supported by a review written by Cunha et al., which listed doxycycline as a preferred agent for treatment of Legionnaire’s disease[13]. Additionally, doxycycline has many potential clinical advantages over azithromycin given its enhanced antibacterial spectrum for many common CAP pathogens. For example, 30.2% of Streptococcus pneumoniae, the most common etiology of CAP, and 13.2% of Mycoplasma pneumoniae are macrolide resistant in the US[14,15]. Another potential benefit of doxycycline is the hypothesized protective effect in the development of Clostridium difficile infection.C. difficile represents a significant source of adverse patient outcomes, with numerous studies demonstrating a variety of antibiotic agents as risk factors for infection[16]. This effect has been demonstrated in a 2012 study in which each day of doxycycline therapy was associated with a 27% lower rate of C. difficile infection as compared to other antimicrobials[17,18].

This study evaluated the efficacy of doxycycline versus azithromycin each in combination with ceftriaxone in veterans admitted to the VA Midsouth Healthcare Network with non-ICU CAP by identifying risk factors associated with a composite endpoint of adverse patient outcomes. These outcomes consisted of the requirement for antibiotic escalation or mortality during the index hospitalization or all-cause readmission, emergency department visit, or mortality within 30 days of discharge. We also evaluated risk factors associated with each component of the composite endpoint in multivariate analyses.


Study design, setting, and patient cohorts

This study was a retrospective, multi-center cohort study of veterans admitted to VA hospitals composing the VA Midsouth Healthcare Network from January 2007 through December 2014. Included veterans were admitted to a general medical ward with a diagnosis of pneumonia per International Classification of Diseases, Ninth Revision code (ICD-9). Patients received ceftriaxone plus either doxycycline or azithromycin within 48 hours of admission. They were placed into either the doxycycline or azithromycin cohort based on the antimicrobial received. Only the first admission per patient during the study period was included as an index hospitalization. Subsequent hospitalizations were included as outcomes if they met study criteria. Veterans could have therapy deescalated at the discretion of the treating physician.

Inclusion and exclusion criteria were based on current IDSA guidelines[5]. Participants were excluded if they had an ICD-9 code indicating a diagnosis of cystic fibrosis, history of tuberculosis, current pregnancy, sepsis, organ transplantation, human immunodeficiency virus infection, or receipt of immunosuppressive therapy (prednisone >15 mg per day, calcineurin inhibitor, mycophenolate, azathioprine, or methotrexate). Patients meeting criteria for Healthcare-Associated Pneumonia (HCAP) were excluded according to previously published guidelines[19]. These criteria included residence in a nursing home or extended care facility, hospitalization for ³48 hours in the preceding 90 days, or presence of end-stage renal disease. Other HCAP criteria, such as home infusion therapy or wound care in the past 30 days or having a family member colonized or infected with a multi-drug resistant pathogen, were not available for collection.

Data collection and study approval

Veterans were selected for inclusion, and data were collected by query of patient level data extracted from the Corporate Data Warehouse database through the VA informatics and Computing Infrastructure. The protocol for research was approved by the applicable institutional review board at the James H. Quillen VA Medical Center.

Demographic data were collected including age, race, and gender. Clinical data, laboratory values, and clinical outcomes were also obtained. The components of the CURB criteria were collected as a measure of pneumonia severity[5]. A modified version of the criteria (mCURB-65) was utilized in this study due to the unavailability of mental status in the Corporate Data Warehouse. Thus, blood urea nitrogen (>19 mg/dL), respiratory rate (>29 breaths per minute), blood pressure (systolic <90 mmHg or diastolic <60 mmHg), and age (³65 years old) were each assigned a score of one if an observed value met criteria <24 hours of admission. Microbiology data, including the presence of positive blood cultures, and baseline comorbidities were collected and assessed as risk factors. Collected baseline comorbidities were utilized to calculate an age-unadjusted Charlson Comorbidity Index[20]. An age un-adjusted value was included due to the presence of age in the mCURB-65 criteria to avoid overfitting of the final model.

Clinical outcomes collected included mortality during admission or death, readmission, or emergency department visit within 30 days of discharge. Escalation of antimicrobial coverage was collected and defined as the receipt of an aminoglycoside (amikacin, gentamicin, tobramycin), carbapenem (ertapenem, imipenem/cilastatin, or meropenem), cefepime, piperacillin/tazobactam, and/or vancomycin after 48 hours of admission. Bar-code scanning data were collected to confirm the antimicrobials received and time of administration.

A random review of charts from the James H. Quillen VA Medical Center was conducted by comparing locally obtained data to information obtained from the Corporate Data Warehouse to ensure data accuracy.


The primary endpoint was a composite outcome consisting of mortality or escalation of antibiotics during the index hospitalization or mortality, hospital readmission, or emergency department visit within 30 days of discharge. Secondary endpoints consisted of the individual components of the primary composite endpoint.

Statistical analysis

Dichotomous variables were compared with c2 or Fishers exact tests as appropriate for their association with the composite outcome in univariate analyses. Variables with a p value <0.20 in univariate analysis were included in multivariate logistic regression analyses to identify independent risk factors associated with the composite outcome. The association of variables with the individual components of the composite outcome was analyzed in secondary analyses by multivariate logistic regression. Odds Ratios (OR) and 95% Confidence Intervals (CI) were calculated and are presented for independent variables. A two-tailed p<0.05 was considered statistically significant. Analyses were conducted using SAS software, version 9.3 (SAS Institute, Cary, NC).


3,788 veterans met inclusion and exclusion criteria for the study with 77 and 3,711 in the doxycycline and azithromycin groups, respectively. Baseline demographics are displayed in table 1. Veterans in the doxycycline group were older with a median age of 73 years compared with 70 years in the azithromycin group. The prevalence of male gender and the racial composition of the two groups were similar with a high percentage of male (98%) Caucasian (84.3%) veterans included in the study. There were numerically more patients with positive blood cultures in the doxycycline group (3.9%) compared with the azithromycin group (1.1%). Legionnaires’ disease was documented in 0.4% of the azithromycin group and not observed in the doxycycline group. Nevertheless, pneumonia severity appeared to be equal between groups with a similar mCURB-65 criterion (median 1 vs 1, respectively). Additionally, veterans in the doxycycline group were numerically more likely to have several comorbidities such as coronary artery disease (29.9% vs 24.7%), heart failure (54.4% vs. 44.4%), and liver cirrhosis (11.7% vs. 3.8%) compared with patients who received azithromycin. The incidence of other comorbidities such as Chronic Obstructive Pulmonary Disease (COPD), asthma, and tobacco use disorder were largely equal between groups. The median CCI was also similar between the two groups (2 vs. 1 in the doxycycline and azithromycin groups, respectively).


Doxycycline (n=77)

Azithromycin (n=3711)

Total (n=3788)

Median age - year (IQR)

73 (65 - 82.5)

70 (62 - 80)

71 (62 - 81)

Male - no. (%)

76 (98.7)

3636 (98.0)

3712 (98.0)

Race - no. (%)




African American

8 (10.4)

356 (9.6)

364 (9.6)

American Indian


6 (0.2)

6 (0.2)


67 (87.0)

3126 (84.2)

3193 (84.3)



29 (0.8)

29 (0.8)

Not available

2 (2.6)

176 (4.7)

178 (4.7)

Asian/Pacific Islander


18 (0.5)

18 (0.5)

Positive blood cultures -no. (%)

3 (3.9)

42 (1.1)

45 (1.2)

Legionnaires’ disease


16 (0.4)

16 (0.4)

Median mCURB-65 (IQR)

1 (1- 2)

1 (1 - 2)

1 (1 - 2)

Median CCI (IQR)

2 (1 - 3)

1 (1 - 3)

1 (1 - 3)

Comorbidities -no. (%)





4 (5.2)

202 (5.4)

206 (5.4)

Coronary artery disease

23 (29.9)

917 (24.7)

940 (24.8)

Chronic obstructive pulmonary disease

47 (61.0)

2383 (64.2)

2430 (64.1)

Cerebrovascular disease

17 (22.1)

755 (20.3)

772 (20.4)

Diabetes Mellitus

33 (42.9)

1529 (41.2)

1562 (41.2)

End-Stage Renal Disease

4 (5.2)

224 (6.0)

228 (6.0)

Heart failure

42 (54.4)

1646 (44.4)

1688 (44.6)

Liver Cirrhosis

9 (11.7)

141 (3.8)

150 (4.0)

Other Respiratory-chronic

3 (3.9)

70 (1.9)

73 (1.9)

Tobacco Use Disorder

27 (35.1)

1301 (35.1)

1328 (35.1)

Table 1: Baseline characteristics.

CCI: Charlson Comorbidity Index; IQR: Interquartile Range; mCURB-65: Modified CURB-65

The clinical outcomes assessed in this study are reported in table 2.51.9% of veterans that received doxycycline met the composite outcome compared with 44.3% of veterans that received azithromycin. Univariate c2 analysis did not identify a statistically significant difference in the incidence of the composite outcome between these two groups (p=0.18).

Primary Outcome

Doxycycline (n=77)

Azithromycin (n=3711)


Composite Outcome

40 (51.9)

1645 (44.3)


Secondary Outcomes




Death within 30 days

4 (5.2)

286 (7.7)


Readmission within 30 days

22 (28.6)

759 (20.5)


Emergency department visit within 30 days

21 (27.3)

800 (21.6)


Antibiotics escalated during index hospitalization

16 (20.8)

615 (16.6)


 Table 2: Outcomes in univariate analyses.

The incidence of the components of the composite outcome for the doxycycline and azithromycin groups are shown in table 2.Mortality during index hospitalization and within 30 days was non-statistically significantly less in the veterans who received doxycycline compared with those who received azithromycin (5.2% vs 7.7%, p=0.41). Doxycycline was associated with a numerically higher rate of all-cause readmission (28.6% vs 20.5%, p=0.08) and emergency department visit (27.3% vs 21.6%, p=0.23) within 30 days and for the percentage of patients that had their antimicrobial coverage escalated during their index hospitalization (20.8% vs 16.6%, p=0.33) compared with veterans who received azithromycin, although none reached statistical significance.

Variables associated with the occurrence of the composite outcome in univariate and multivariate analyses are displayed in table 3.Receipt of doxycycline met the prespecified cutoff (p<0.20) in univariate analysis (p=0.18); thus, it was included in the multivariate model. It was not found to be a predictor of the composite outcome in multivariate logistic regression (p=0.26). Male gender was not predictive of the composite outcome in univariate analysis (p=0.78).However, the mCURB-65 [2 vs 0] (OR 1.24: 95%CI 1.05 - 1.47, p=0.01), mCURB-65 [>3 vs 0] (OR 2.40: 95% CI 1.22 - 4.71, p=0.01), CCI (OR 1.10: 95%CI 1.06 - 1.14, p<0.0001), and presence of a positive blood culture (OR 5.81: 95%CI 2.69 - 12.55, p<0.0001) were demonstrated to be statistically significant predictors of the composite outcome.


No composite outcome (n=2103)

Composite outcome (n=1685)


p Value


p value

Odds ratio

(95% CI)

Received doxycycline - no. (%)

37 (1.8)

40 (2.4)




Male - no. (%)

2062 (98.1)

1650 (97.8)




Positive blood culture

8 (0.4)

37 (2.2)



5.81 (2.69 - 12.55)

Median mCURB-65 (IQR)

1 (0 - 2)

1 (1 - 2)




1 vs 0






2 vs 0





1.24 (1.05 - 1.47)

3 vs 0





2.40 (1.22 - 4.71)

Median CCI (IQR)

1 (1 - 3)

3 (1 - 3)



1.10 (1.06 - 1.14)

Table 3:Risk factors for composite outcome.

CCI: Charlson Comorbidity Index; IQR: Interquartile Range; mCURB-65: Modified CURB-65

Table 4displays results from secondary multivariate analyses conducted to determine predictors of the components of the primary composite outcome. Receipt of doxycycline was not found to be a statistically significant risk factor for any of the individual components of the composite outcome. Presence of a positive blood culture was demonstrated to be a predictor of several outcomes, including escalation of antibiotics during index hospitalization and death during or 30 days after hospitalization. The mCURB-65 was found to be a predictor of mortality during the index hospitalization or within 30 days of discharge, an emergency department visit within 30 days, and antibiotic escalation during the index hospitalization. The CCI was statistically associated with several of the outcomes, including death during hospitalization or within 30 days of discharge, readmission within 30 days, and an emergency department visit within 30 days.

Risk factors for readmission within 30 days
Parameter No outcome (n=3007)

Outcome (n=781)

p value
Odds ratio
(95% CI)
Received doxycycline - no. (%) 55 (1.8) 22 (2.8) 0.99  
Positive blood culture 35 (1.2) 10 (1.3) 0.87  
Median mCURB-65 (IQR) 1 (1 - 2) 1 (1 - 2)    
1 vs 0     0.36  
2 vs 0     0.24  
≥3 vs 0     0.84  
Median CCI (IQR) 1 (1 - 3) 3 (1 - 3) <0.0001 1.09 (1.05 - 1.14)


Risk factors for death within 30 days
Parameter No outcome
p value
Odds ratio
(95% CI)
Received doxycycline -no. (%) 73 (2.1) 4 (1.4) 0.26  
Positive blood culture 30 (0.9) 15 (5.2) <0.0001 6.58 (3.45 - 12.55)
Median mCURB-65 (IQR) 1 (0 - 2) 2 (1 - 2)    
1 vs 0     0.02 1.60 (1.09 - 2.37)
2 vs 0     <0.0001 2.43 (1.68 - 3.52)
≥3 vs 0     <0.0001 5.95 (2.54 - 13.96)
Median CCI (IQR) 1 (1 - 3) 3 (1 - 3) <0.0001 1.16 (1.10- 1.23)


Risk factors for emergency department visit within 30 days
Parameter No outcome
p value
Odds ratio
(95% CI)
Received doxycycline -no. (%) 56 (1.9) 21 (2.6) 0.27  
Positive blood culture 38 (1.3) 7 (0.9) 0.29  
Median mCURB-65 (IQR) 1 (0 - 2) 1 (1 - 2)    
1 vs 0     0.01 1.32 (1.07 - 1.63)
2 vs 0     0.02 1.28 (1.04 - 1.58)
≥3 vs 0     0.12  
Median CCI (IQR) 1 (1 - 3) 3 (1 - 3) 0.03 1.05 (1.01 -1.09)


Risk factors for antibiotic escalation during index hospitalization
Parameter No outcome
p value
Odds ratio
(95% CI)
Received doxycycline -no. (%) 61 (1.9) 16 (2.5) 0.50  
Positive blood culture 14 (0.4) 31 (4.9) <0.0001 11.56 (6.11 - 21.88)
Median mCURB-65 (IQR) 1 (1 - 2) 1 (1 - 2)    
1 vs 0     0.84  
2 vs 0     0.25  
≥3 vs 0     0.01 2.56 (1.60- 5.19)
Median CCI (IQR) 1 (1 - 3) 2 (1 - 3) 0.94  

Table 4:Secondary multivariate analyses.

CCI: Charlson Comorbidity Index; IQR: Interquartile Range; mCURB-65: Modified CURB-65


In the present study, no statistically significant differences in patient outcomes were observed between regimens containing doxycycline and azithromycin for the treatment of non-ICU CAP. The two cohort groups were similar regarding pneumonia severity, with modified CURB criteria comparable between groups. Veterans included in the doxycycline group had a higher rate of positive blood cultures; the reasons for this increase in positive blood cultures are not readily apparent. Of note, organism identification was not available for analysis; thus, the percentages of contaminants are not known. However, the percentages are likely equal between groups due to similar overall patient demographics, obtainment by the same personnel, and processing in the same microbiology laboratories.CCI scores were slightly higher in the doxycycline group compared with the azithromycin group (median 2 vs 1, respectively), with select comorbidities also numerically higher (Table 1). The incidence of cardiovascular disease in the doxycycline group was numerically greater than the azithromycin group, likely indicating preferential use of this agent in patients perceived to be at higher cardiovascular risk.

The numerical increase in the composite outcome observed in the doxycycline group appears to be driven by the larger percentage of readmissions and emergency department visits. These increases may be due to a higher proportion of select comorbidities in veterans that received doxycycline. These veterans had higher rates of heart failure, coronary artery disease, and liver cirrhosis, which may have predisposed these patients to an increased risk of all-cause emergency department visits or all-cause readmission. There were slightly more patients diagnosed with COPD in the azithromycin group (absolute difference 3.2% between groups), likely a result of azithromycins longstanding use in this patient population. Notably, the rate of mortality during index hospitalization and within 30 days was not statistically different in the doxycycline group and was numerically greater for veterans that received azithromycin. This study was not powered to detect a statistical difference for this endpoint; however, the similar rates for this important clinical outcome are notable. Additionally, 30-day mortality rates observed in this study (7.6%) appear similar to other studies that reported mortality rates of 6.5% to 12.1% depending on the severity of pneumonia and patient population studied[3,21-23]. Secondary analyses identified several predictors for the different components of the composite outcome (Table 4). These included markers of pneumonia severity and measures of baseline comorbidities. Doxycycline was not observed to be a significant predictor of any of the secondary outcomes.

This study is the first to directly compare ceftriaxone in combination with either doxycycline or azithromycin in non-ICU CAP. Several similar studies have been conducted; however, several important limitations exist for each[24-27]. The closest comparator article is an analysis conducted by Teh et al. [25]. This study compared patients recruited to the Australian Community-Acquired Pneumonia Study, a prospective, multi-center, observational trial[28]. Patients received a beta-lactam plus either doxycycline or a macrolide. Teh et al., reported a shorter time to clinical stability [2 days (0 - 7) vs. 2 days (0 - 22), p=0.006] and length of stay [5 days (0 -26) vs 6 days (0 - 78), p<0.001] for patients in the doxycycline group versus macrolide group, respectively[25].However, several important limitations must be taken into consideration when evaluating these results. The antimicrobials provided appear to be heterogeneous, with only 36% of patients receiving ceftriaxone as their beta-lactam component and the majority receiving benzylpenicillin or an aminopenicillin. Additionally, numerous macrolides were permitted, with some patients receiving roxithromycin which is not available in the US. An additional limitation of this trial is a result of the observational nature of the treatment assignment, with a greater percentage of patients in the macrolide group requiring vasopressors or ventilator support. The authors note that this may have resulted in significant confounding[25].

Flanders et al., conducted a retrospective cohort study of hospitalized adults who received either ceftriaxone and doxycycline or author-defined other appropriate regimens[24].They reported a reduced inpatient mortality [2.3% vs. 14.4%, OR 0.26: 95% CI (0.08 - 0.81)] and 30-day mortality [6.0% vs 20%, OR 0.37: 95% CI (0.17-0.81)] with no difference in length of stay or readmission for the doxycycline group versus other appropriate therapy group, respectively.Up to 29% of patients in the comparator group may have received a beta-lactam and doxycycline containing regimen, with 42% of antibiotic usage unreported. Further complicating interpretation of this trial were the numerous differences in baseline characteristics, with increased pneumonia severity, comorbid illness, admission to ICU, aspiration, and admission from long-term care facilities in the other appropriate therapy group. The authors utilized a propensity score to adjust for the differences in baseline characteristics but note that the results of the trial could be due to residual confounding.

Other studies reported in the literature investigating the use of doxycycline for inpatient CAP utilized it as monotherapy[26,27]. Doxycycline was shown to have a decreased time to clinical response, shorter length of stay, and decreased cost compared with an unreported provider-selected regimen[27].In an analysis versus levofloxacin, doxycycline was associated with a decreased length of stay and cost with a similar rate of failure[26]. Current clinical evidence for this regimen provided in the 2007 IDSA/ATS guideline is limited to two studies published in 1997 and 1999, and both evaluated doxycycline as monotherapy[27,29].

CAP remains a significant source of morbidity and mortality in various patient populations; however, this is particularly true for patients with immunocompromising conditions and elderly patients. Patients older than 65 years of age are 4.17 times as likely to develop CAP than patients younger than 45 years of age[30]. A large US study demonstrated a greater incidence of hospitalization for CAP with patients 65 to 79 years and patients over 80 years of age experiencing rates 9 and 25 times higher than patients 18 to 40 years of age, respectively[31]. These patients at increased risk of deleterious outcomes from CAP may also be at higher risk of azithromycin cardiotoxicity. An analysis in 2012 reported a linear increase in cardiovascular death with the decile of a cardiovascular risk score[6]. The increased cardiovascular toxicity of azithromycin has been subsequently evaluated in numerous publications with some confirming[10,12]and others finding either no increase or an overall mortality benefit[7,9]. The VA population contains a high percentage of patients at elevated cardiovascular risk. Thus, clinicians caring for patients at high baseline risk may want to consider alternative therapies such as doxycycline.

Antibiotic decisions for CAP are frequently made prior to obtainment of microbiological data, necessitating the use of empiric guideline recommendations and local epidemiologic data. Furthermore, the incidence of CAP cases in which an etiology is not determined has been shown to range from 55% to 66%[32]. Fluoroquinolones are also recommended for the empiric treatment of non-ICU inpatient CAP. However, these agents are generally not preferred in the VA patient population due to recent FDA warnings regarding adverse effects, concerns of increasing antimicrobial resistance, and their impact on facilitating C. difficile acquisition[33].

Several important limitations must be considered when analyzing the results of this study. This was a retrospective observational study; thus, confounding may have occurred with regard to regimen selection by the treating physician.HCAP was removed from the recent hospital-acquired and ventilator-associated pneumonia guidelines[34]. These criteria were utilized as exclusion criteria despite numerous studies demonstrating their poor predictive ability to identify patients at risk of resistant pathogens[5,35-37]. Veterans meeting criteria for HCAP were excluded to enable analysis of a more homogenous sample and due to the clinical uncertainty with regard to treating these patients. The absolute number of patients in the doxycycline group were low in comparison to the azithromycin group. This is likely due to the preferential utilization of azithromycin over doxycycline in accordance with guidelines and the potential presence of increased exclusion criteria in patients that were prescribed doxycycline. Due to the small numbers, a power calculation was not performed, and the presence of a type II error cannot be excluded.

The unavailability of mental status on admission is a limitation of the study. However, the potential confounding on the results of the trial are minimal since this information was not available for either group, and it appears that the two groups were well-matched in regard to the modified CURB criteria. Also, the low CURB score indicates that many of these patients may have qualified for outpatient treatment, indicating a lower risk of expected adverse outcomes. However, we observed a high rate of patients meeting the composite outcome. One explanation for this may be the high rate of comorbidities present in the VA patient population.


The administration of doxycycline as an alternative to azithromycin in the treatment of inpatient non-ICU CAP was not associated with a statistical difference in a composite endpoint of negative patient outcomes. This study is the first to our knowledge to directly compare doxycycline and azithromycin in this patient population. While there was an appreciable difference in study group sizes, our data supports current IDSA/ATS guideline recommendations regarding the use of doxycycline as an alternative to azithromycin as part of combination therapy with ceftriaxone. These data are especially encouraging for patients with a macrolide allergy or considered to be at high baseline risk of cardiovascular adverse effects. Given contradictory literature and Food and Drug Administration warnings of azithromycin cardiovascular toxicity, doxycycline may represent an attractive option for many patients with this common infectious disease.


This publication is the result of work supported with resources and the use of facilities at the Mountain Home VA Healthcare System. The contents in this publication do not represent the views of the Department of Veterans Affairs or theUnited States Government.


  1. Kochanek KD, Murphy SL, Xu J, Tejada-Vera B (2016) Deaths: Final Data for 2014.Natl Vital Stat Rep65: 1-122.
  2. Ramirez JA, Wiemken TL, Peyrani P, Arnold FW, Kelley R, et al. (2017) Adults Hospitalized with Pneumonia in the United States: Incidence, Epidemiology & Mortality. Clin Infect Dis 65: 1806-1812.
  3. Ito A, Ishida T, Tokumasu H, Washio Y, Yamazaki A, et al. (2017) Prognostic factors in hospitalized community-acquired pneumonia: a retrospective study of a prospective observational cohort. BMC Pulm Med 1: 78.
  4. McLaughlin JM, Johnson MH, Kagan SA, Baer SL (2015) Clinical and economic burden of community-acquired pneumonia in the Veterans Health Administration, 2011: a retrospective cohort study. Infection 6: 671-680.
  5. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, et al. (2007) Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect 1: 27-72.
  6. Ray WA, Murray KT, Hall K, Arbogast PG, Stein CM (2012) Azithromycin and the risk of cardiovascular death. N Engl J Med 20: 1881-1990.
  7. Svanstrom H, Pasternak B, Hviid A (2013) Use of azithromycin and death from cardiovascular causes. N Engl J Med 18: 1704-1712.
  8. Cheng YJ, Nie XY, Chen XM, Lin XX, Tang K, et al. (2015) The Role of Macrolide Antibiotics in Increasing Cardiovascular Risk. J Am Coll Cardiol 20: 2173-2184.
  9. Mortensen EM, Halm EA, Pugh MJ, Copeland LA, Metersky M, et al. (2014) Association of azithromycin with mortality and cardiovascular events among older patients hospitalized with pneumonia. JAMA 21: 2199-2208.
  10. Rao GA, Mann JR, Shoaibi A, Bennett CL, Nahhas G, et al. (2014) Azithromycin and levofloxacin use and increased risk of cardiac arrhythmia and death. Ann Fam Med 2: 121-127.
  11. Trac MH, McArthur E, Jandoc R, Dixon SN, Nash DM, et al. (2016) Macrolide antibiotics and the risk of ventricular arrhythmia in older adults. CMAJ 7: 120-129.
  12. Chou HW, Wang JL, Chang CH, Lai CL, Lai MS, et al. (2015) Risks of cardiac arrhythmia and mortality among patients using new-generation macrolides, fluoroquinolones, and beta-lactam/beta-lactamase inhibitors: a Taiwanese nationwide study. Clin Infect Dis 4: 566-577.
  13. Cunha CB, Cunha BA (2017) Antimicrobial Therapy for Legionnaire’s Disease: Antibiotic Stewardship Implications. Infect Dis Clin North Am 1: 179-191.
  14. Kim L, McGee L, Tomczyk S, Beall B (2016) Biological and Epidemiological Features of Antibiotic-Resistant Streptococcus pneumoniae in Pre- and Post-Conjugate Vaccine Eras: a United States Perspective. Clin Microbiol Rev 3: 525-352.
  15. Zheng X, Lee S, Selvarangan R, Qin X, Tang YW, et al. (2015) Macrolide-Resistant Mycoplasma pneumoniae, United States. Emerg Infect Dis 8: 1470-1472.
  16. Cohen SH, Gerding DN, Johnson S, Kelly CP, Loo VG, et al. (2010)Clinical practice guidelines for Clostridium difficile infection in adults: 2010 update by the society for healthcare epidemiology of America (SHEA) and the infectious diseases society of America (IDSA). Infect Control Hosp Epidemiol 5: 431-455.
  17. Turner RB, Smith CB, Martello JL, Slain D (2014) Role of doxycycline in Clostridium difficile infection acquisition. Ann Pharmacother 48: 772-776.
  18. Doernberg SB, Winston LG, Deck DH, Chambers HF (2012) Does doxycycline protect against development of Clostridium difficile infection? Clin Infect Dis 5: 615-620.
  19. AmericanThoracic Society; Infectious Diseases Society of America(2005) Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 4: 388-416.
  20. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 5: 373-383.
  21. Musher DM, Thorner AR (2015) Community-acquired pneumonia. N Engl J Med 3: 294.
  22. Lee JS, Nsa W, Hausmann LR, Trivedi AN, Bratzler DW, et al. (2004) Quality of care for elderly patients hospitalized for pneumonia in the United States, 2006 to 2010. JAMA Intern Med 11: 1806-1814.
  23. Metersky ML, Waterer G, Nsa W, Bratzler DW (2012) Predictors of in-hospital vs postdischarge mortality in pneumonia. Chest 2: 476-481.
  24. Flanders SA, Dudas V, Kerr K, McCulloch CE, Gonzales R (2006) Effectiveness of ceftriaxone plus doxycycline in the treatment of patients hospitalized with community-acquired pneumonia. J Hosp Med 1: 7-12.
  25. Teh B, Grayson ML, Johnson PD, Charles PG (2012) Doxycycline vs. macrolides in combination therapy for treatment of community-acquired pneumonia. Clin Microbiol Infect 4: 71-73.
  26. Mokabberi R, Haftbaradaran A, Ravakhah K (2010) Doxycycline vs. levofloxacin in the treatment of community-acquired pneumonia. J Clin Pharm The 2: 195-200.
  27. Ailani RK, Agastya G, Ailani RK, Mukunda BN, Shekar R (1999) Doxycycline is a cost-effective therapy for hospitalized patients with community-acquired pneumonia. Arch Intern Med 3: 266-270.
  28. Charles PG, Whitby M, Fuller AJ, Stirling R, Wright AA, et al. (2008) The etiology of community-acquired pneumonia in Australia: why penicillin plus doxycycline or a macrolide is the most appropriate therapy. Clin Infect Dis 10: 1513-1521.
  29. Ragnar Norrby S (1997) Atypical pneumonia in the Nordic countries: aetiology and clinical results of a trial comparing fleroxacin and doxycycline. Nordic Atypical Pneumonia Study Group. J Antimicrob Chemother 4: 499-508.
  30. Baik I, Curhan GC, Rimm EB, Bendich A, Willett WC, et al. (2000) A prospective study of age and lifestyle factors in relation to community-acquired pneumonia in US men and women. Arch Intern Med 20: 3082-3088.
  31. Jain S, Self WH, Wunderink RG (2015) Community-Acquired Pneumonia Requiring Hospitalization among US. Adults. N Engl J Med 5: 415-427.
  32. Musher DM, Abers MS, Bartlett JG (2017) Evolving Understanding of the Causes of Pneumonia in Adults, With Special Attention to the Role of Pneumococcus. Clin Infect Dis 10: 1736-1744.
  33. Leffler DA, Lamont JT (2015) Clostridium difficile Infection. N Engl J Med 3: 287-288.
  34. Kalil AC, Metersky ML, Klompas M, Muscedere J, Sweeney DA, et al. (2016) Management of Adults With Hospital-acquired and Ventilator-associated Pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis 5: 61-111.
  35. Chalmers JD, Rother C, Salih W, Ewig S (2014)Healthcare-associated pneumonia does not accurately identify potentially resistant pathogens: a systematic review and meta-analysis. Clin Infect Dis 3: 330-339.
  36. Gross AE, Van Schooneveld TC, Olsen KM, Rupp ME, Bui TH, et al. (2014) Epidemiology and predictors of multidrug-resistant community-acquired and healthcare-associated pneumonia. Antimicrob Agents Chemother 9: 5262-5268.
  37. Chalmers JD, Taylor JK, Singanayagam A, Fleming GB, Akram AR, et al. (2011) Epidemiology, antibiotic therapy, and clinical outcomes in healthcare-associated pneumonia: a UK cohort study. Clin Infect Dis 2: 107-113.

Key Points

  • Doxycycline is recommended as an alternative to azithromycin for patients with CAP based on limited clinical evidence and in vitro data.
  • In addition to those with a macrolide allergy or significant intolerance, patients who have considerable cardiovascular comorbidities are often prescribed doxycycline to circumvent any issues.
  • In this targeted study, doxycycline performed similarly to azithromycin which bolsters current IDSA and ATS recommendations.
Ocimum Scientific Publishers

This work is licensed under a Creative Commons Attribution 4.0 International License.  Creative Commons License

Copyright © 2019 - All Rights Reserved -