Antipseudomonal Monotherapy or Combination Therapy for Older Adults with Community-Onset Pneumonia and Multidrug Resistant Risk Factors: A Retrospective Cohort Study

Abstract

Background:

Infectious Diseases Society of America (IDSA) guidelines recommend empiric antipseudomonal combination therapy when Pseudomonas is suspected. However, combination antipseudomonal therapy is controversial. This study compares all-cause 30-day mortality in older patients who received antipseudomonal monotherapy (PMT) or antipseudomonal combination therapy (PCT) for the treatment of community-onset pneumonia.

Methods:

This population-based, retrospective, cohort study used data from over 150 Veteran Health Administration hospitals. Patients were classified as low, medium, or high risk of drug-resistant pathogens. 31,027 patients were assigned to PCT or PMT treatment arms based on antibiotics received in the first 48 hours of hospital admission.

Results:

The unadjusted 30-day mortality difference between PCT and PMT was most pronounced in the low-risk group (18% vs. 8%), followed by the medium-risk group (24% vs. 18%) and the high-risk group (39% vs. 33%). PCT was associated with higher 30-day mortality than PMT overall (aOR, 1.54; 95%CI, 1.43-1.66), in all three risk groups: low (aOR, 1.69; 95%CI, 1.50-1.89), medium (aOR, 1.30; 95%CI, 1.14-1.48), and high (aOR, 1.21; 95%CI, 1.04-1.40).

Conclusion:

Older adults who received combination antipseudomonal therapy for community-onset pneumonia fared worse than those who received monotherapy. Empiric combination antipseudomonal therapy should not be routinely offered to all patients suspected of having pseudomonal pneumonia.

Background

Antibiotics are the mainstay of treatment for pneumonia; however, therapy is not “one size fits all”. Furthermore, timely administration of antibiotics is critical, so initial empiric treatment, followed by pathogen-directed therapy, is required. One of the most controversial questions to date involves the choice of empiric antipseudomonal monotherapy (PMT) or antipseudomonal combination therapy (PCT) for patients suspected of having community-onset pneumonia due to Pseudomonas. Clinicians want to do what is best for their patients, including adequate coverage, to increase the probability of a positive outcome; however, they also want to limit their patients’ exposure to broad-spectrum antibiotic therapies that may result in health complications and promote antibiotic resistance. Antibiotics are a precious resource and prescribers want to be good stewards, but lack high-quality evidence to inform this important decision. The available evidence comes from small, single-center studies, and is often contradictory [1]. Clinicians are left to wonder, “Should I routinely prescribe PCT for my patients presenting with community-onset pneumonia and multidrug resistant (MDR) risk factors?”

When Pseudomonas pneumonia is suspected, American Thoracic Society / Infectious Diseases Society of America (ATS/IDSA) pneumonia guidelines recommend PCT with two or more antipseudomonal therapies, including: beta-lactam plus aminoglycoside, beta-lactam plus fluoroquinolone, or aminoglycoside plus fluoroquinolone [2]. Nevertheless, several studies have failed to observe additional benefit with PCT versus PMT in the setting of Pseudomonas pneumonia [3–7]. Early studies were promising for PCT [8], but more recent studies have not been able to reliably confirm those findings [9–11]. Additional studies have demonstrated a mortality benefit only for the sickest pneumonia patients [12], including those with Pseudomonas bacteremia or as definitive treatment [13,14].

Since clinicians rarely know if a patient has Pseudomonas bacteremia on admission, and timely empiric therapy is critical, it is helpful to use prediction rules to decide who gets any type of empiric antipseudomonal therapy (PMT or PCT). One of our prior studies confirmed that such rules can be used effectively to identify patients likely to benefit from empiric antipseudmonal therapy [15]. However, it is unclear if such rules can also be used to identify patients who might experience additional benefit from PCT (versus PMT); therefore, the objective of this study was to compare all-cause 30-day mortality in older patients who received PCT (versus PMT) for the treatment of community-onset pneumonia, stratified based on risk level of having pseudomonal pneumonia.

Methods

This population-based, retrospective, cohort study used data from over 150 hospitals and 1,400 clinics in the Veterans Health Administration (VHA) system between fiscal years 2002 and 2007. The methods used to build the database, and define the study population, have been previously published [16–19]. Briefly, data for this study were obtained from the VHA electronic medical record system that includes administrative, clinical, laboratory, and pharmacy data. These data were generated during the course of care and billing for VHA patients, and not for research; however, the study investigators used these pieces of data to construct variables that can be used for research. For instance, comorbidities were defined using International Classification of Diseases (ICD) codes. It is important to note that this was not a sample or subset of the VHA population. Rather, this was a population-based study of all patients in the VHA system. The Institutional Review Board of the University of Texas Health Science Center at San Antonio and the South Texas Veterans Health Care System Research and Development committee approved this study.

Patients were included if they were ≥ 65 years of age and had either a primary discharge diagnosis of pneumonia/influenza (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 480.0–483.99 or 485–487) or a secondary discharge diagnosis of pneumonia/influenza plus a primary diagnosis of respiratory failure (ICD-9-CM code 518.81) or sepsis (ICD-9-CM code 038.xx). Comorbid conditions were determined using ICD-9-CM codes from outpatient and inpatient care in accordance with the Charlson comorbidity scoring system [20, 21]. If patients were admitted more than once during the period of study, only the first episode qualified for inclusion in the study. Patients who did not receive antipseudomonal therapy, and those who received it after 48 hours post admission, were excluded.

The risk score tool and variables for the stratification of these patients was based on the risk score developed by Ma et al. and modified for our database, as earlier defined [17]. Briefly, the risk level groups [low (0.0-2.0), medium (2.5-7.0), and high (7.5-21.5)], were generated based on the number of points a patient had, as contributed by the total score from the comorbidities presented by the individual. A total of 31,027 patients received antipseudomonal therapy in the first 48 hours of admission and were stratified into PMT or PCT arms for each risk group. Antipseudomonal therapies were defined as the receipt of specific beta-lactams, fluoroquinolones, or aminoglycosides with in vitro activity against Pseudomonas. Table 1 depicts the complete list of all antibiotics considered in this study, and definitions of guideline-concordant community-acquired pneumonia (GC-CAP) therapy, Pseudomonas therapy, and methicillin-resistant Staphylococcus aureus (MRSA) therapy.

Table 1.

Definitions of antibiotic therapy

Guideline-Concordant CAP therapy
Ward patients  • Beta-lactam1 plus (macrolide2 or doxycycline)  • Respiratory fluoroquinolone3	ICU patients  • Beta-lactam1 plus (macrolide2 or doxycycline)  • Beta-lactam1 plus respiratory fluoroquinolone3
Pseudomonas therapy
• Antipseudomonal beta-lactam4  • Antipseudomonal fluoroquinolone5  • Aminoglycoside6
MRSA therapy
• Vancomycin  • Linezolid

CAP: community-acquired pneumonia; ICU: intensive care unit

¹Beta-lactam includes cefotaxime, ceftriaxone, ampicillin-sulbactam, ertapenem, or aztreonam

²Macrolide includes azithromycin, clarithromycin, or erythromycin

³Respiratory fluoroquinolone includes moxifloxacin, levofloxacin, orgatifloxacin

⁴Antipseudomonal beta-lactam includes cefepime, ceftazidime, imipenem-cilastatin, meropenem, piperacillin-tazobactam, ticarcillin-clavulanate, or aztreonam

⁵Antipseudomonal fluoroquinolone includes ciprofloxacin or levofloxacin

⁶Aminoglycoside includes gentamicin, tobramycin, or amikacin

The primary study outcome was all-cause 30-day mortality, which has been shown to be most closely associated with pneumonia-related mortality [22]. Mortality was assessed using the VHA Vital Status file, which is in close agreement (98%) with the National Death Index [23].

Statistical Analyses

30-day all-cause mortality was compared for patients who received PMT or PCT. All statistical analyses were conducted using JMP 10.0^® (SAS Corp, Cary, NC). Categorical variables were compared by Fisher’s exact test or Chi-square. Continuous variables were compared using the Wilcoxon Rank Sum test (Table 2). For bivariable statistical tests, P values of ≤ 0.05 were considered to be statistically significant, and all tests were two-tailed.

Table 2.

Baseline characteristics for patients with antipseudomonal monotherapy (PMT) or antipseudomonal combination therapy (PCT)

	PMT (n=23,916)	PCT (n=7,111)	P-value
Patient age (years), median (IQR)	78 (72-82)	78 (72-83)	0.39
Male, %	98	99	0.0093
Race, %
White	83	80	0.0002
Black	12	13	0.70
Other	5	7	<0.0001
Hispanic ethnicity, %	6	10	<0.0001
MDR risk score variables (points), %
Respiratory organ failure (14pts)	8	21	<0.0001
Hospitalization in the past 90 days (5pts)	24	38	<0.0001
Invasive mechanical ventilation (2pts)	4	17	<0.0001
Healthcare-associated pneumonia risk factor (0.5pts)	35	50	<0.0001
MDR risk score, median (IQR)	0 (0-5.5)	5.5 (0-5.5)	<0.0001
Low (0.0-2.0), %	70	48	<0.0001
Medium (2.5-7.0), %	21	30	<0.0001
High (7.5-21.5), %	9	22	<0.0001
Charlson comorbidity score, median (IQR)	2 (1-4)	3 (1-4)	<0.0001
Comorbid conditions, %
Myocardial infarction	7	8	<0.0001
Heart failure	26	26	0.86
Chronic obstructive pulmonary disease	53	52	0.0097
Liver disease	1	1	0.17
Renal disease	12	15	<0.0001
Diabetes	33	35	0.0039
Neoplastic disease	26	30	<0.0001
HIV/AIDS	<1	<1	0.13
Medication use within 90 days, %
Cardiovascular medications	71	65	<0.0001
Anti-diabetic medications	23	22	0.53
Inhaled corticosteroids	24	21	<0.0001
Systemic corticosteroidsa	24	25	0.37
Pulmonary medications	39	36	<0.0001
Vasopressors, %	3	12	<0.0001
Invasive mechanical ventilation, %	4	17	<0.0001
Noninvasive mechanical ventilation, %	3	7	<0.0001
Hemodialysis, %	15	20	<0.0001
Organ failure, %
Any organ failure, %	21	39	<0.0001
Respiratory	8	21	<0.0001
Cardiovascular	5	9	<0.0001
Neurological	2	3	<0.0001
Renal	11	21	<0.0001
Hematologic	2	5	<0.0001
Hepatic	<1	<1	0.0012
Antibiotic therapy, %
Guideline-concordant CAP therapy	85	67	<0.0001
MRSA therapy	49	59	<0.0001
Pseudomonas diagnosis code by discharge	1 (179)	7 (504)	<0.0001

CAP: community-acquired pneumonia; HIV/AIDS: human immunodeficiency virus/acquired immunodeficiency syndrome; IQR: interquartile range; MRSA: methicillin-resistant Staphylococcus aureus; PSE: Pseudomonas

^aIncludes oral and/or injectable corticosteroids

Potential confounding was minimized using multivariate logistic regression. Separate multivariable logistic regression models were constructed to determine whether any association existed between PMT or PCT and 30-day mortality in the overall population and additionally in each of the three risk groups. The dependent variable was 30-day mortality, and the independent variable was PMT or PCT. Covariates included all characteristics from Table 2. The independent variable and all covariates were entered into the model simultaneously. Adjusted odds ratios (aORs) and 95% confidence intervals (95%CI) were calculated; those 95%CI not crossing 1 were considered to be statistically significant.

The final list of covariates included: patient age, race, Hispanic ethnicity, myocardial infarction, heart failure, chronic obstructive pulmonary disease, liver disease, renal disease, diabetes, neoplastic disease, cardiovascular medications, anti-diabetic medications, inhaled corticosteroids, systemic corticosteroids, pulmonary medications, vasopressors, invasive and non-invasive mechanical ventilation, respiratory failure, cardiovascular failure, neurological failure, renal failure, hematological failure, hepatic failure, GC-CAP therapy, and MRSA therapy. When variables were collinear, only one of the variables was used. For instance, most patients on hemodialysis also had renal failure; therefore, renal failure was chosen as the variable for the models and the hemodialysis variable was excluded. The Charlson score and the “any organ failure” variables were excluded from the models because individual comorbidities and organ failures were already included in the models. Individual risk score variables were also excluded from the models because our study ran separate multivariable models for the three risk groups, and these individual characteristics were used to define those risk groups.

Results

Overall Population

Of the 31,027 patients who met study criteria, 23% received PCT and 77% received PMT. Patients belonged to low (59%), medium (24%), and high (18%) risk groups. One or more healthcare-associated pneumonia (HCAP) criteria was the most common Multi Drug Resistant (MDR) risk score variable, followed by hospitalization in the past 90 days, respiratory organ failure, and invasive mechanical ventilation.

The median (interquartile range) Charlson score was 2 (1-4) and common comorbidities included chronic obstructive pulmonary disease (53%), diabetes (33%), heart failure (26%), and neoplastic disease (26%). The most common medications used within 90 days prior to admission were cardiovascular (71%) and pulmonary (39%) medications. Organ failure occurred in 24% of patients. Finally, 80% of patients received GC-CAP therapy and 34% received MRSA therapy.

Baseline Characteristics

Patient age (median of 78 years), race (>80% white), and sex (>98% male) were similar for patients receiving PCT or PMT (Table 2); however, many comorbid conditions including diabetes, renal disease, and heart failure were more prevalent in the PCT arm. All the MDR risk score variables were more prevalent in the PCT arm, and a greater proportion of patients in the PCT arm had organ failure.

Of note, 70% of those on PMT were classified as low-risk, compared to 50% of those on PCT. Conversely, 22% of those on PCT were high-risk, as compared to 9% of those on PMT. Finally, PMT patients were more likely to have received guideline-concordant therapy, whereas PCT patients were more likely to have received MRSA therapy and to have had a positive Pseudomonas culture by discharge.

All-cause Mortality

30-day mortality was 18% overall and increased among the risk groups: low (13%), medium (21%), and high (36%). The unadjusted mortality difference between PCT and PMT was most pronounced in the low-risk group (18% vs. 8%, 10% absolute risk difference), followed by the medium-risk group (24% vs. 18%, 6% difference) and the high-risk group (39% vs. 33%, 6% difference) (Figure 1). PCT was associated with higher 30-day mortality than PMT overall (aOR, 1.54; 95%CI, 1.43-1.66), and in all three risk groups: low (aOR, 1.69; 95%CI, 1.50-1.89), medium (aOR, 1.30; 95%CI, 1.14-1.48), and high (aOR, 1.21; 95%CI, 1.04-1.40).

Figure 1.

Comparison of 30-day mortality in patients who received antipseudomonal monotherapy (PMT) and antipseudomonal combination therapy (PCT)

aOR: adjusted odds ratio; 95% CI: 95% confidence interval

Discussion

The present study set out to determine if PCT (versus PMT) is associated with additional survival benefit in the high-risk group. Unfortunately, the present study found no additional benefit with PCT in any of the risk groups, including the high-risk group. Our results favored PMT significantly, with 33% mortality in the monotherapy arm versus 39% in the combination therapy arm (aOR 1.21, 95% CI 1.04 – 1.40).

Other studies have also demonstrated a lack of additional benefit with combination therapy versus monotherapy in community-onset pneumonia and other infections [9, 11, 24]. A 2004 systematic review evaluating beta-lactam monotherapy versus beta-lactam plus aminoglycoside combination therapy for sepsis found no advantage to combination therapy among patients with gram-negative infections including Pseudomonas aeruginosa infections [9]. The authors reported that patients with pneumonia had significantly fewer failures with monotherapy and there was no advantage to combination therapy. In another study, combination therapy was not associated with improved survival in patients with Gram-negative bacteremia (odds ratio, 0.96; 95%CI, 0 .70-1.32), [11] possibly due to inadequate sample size. In another study, hospital mortality was numerically higher for PCT patients (37%) compared with PMT patients (29%) (p=0.17) [24]. Though not statistically significant, the findings suggest that PCT might have been associated with poorer outcome than PMT had the study authors included a larger sample size. Several additional studies have reported no survival benefit for empiric PCT versus PMT [25–28]. The mortality rates in our study are greater than most of the prior studies, likely because our population is older with a greater proportion of patients who are male and have serious comorbidities. Nevertheless, the association observed in our study is consistent with that observed in prior studies. Furthermore, because of our large sample size, we are able to not only observe the same association (i.e., numerically better survival with PMT vs. PCT), but also observe a statistically significant difference, in favor of PMT over PCT. While our study is not the first to demonstrate this lack of additional benefit with PCT versus PMT, it is the first to do so among high-risk patients—the group that is considered to most likely benefit from PCT, and a group that has previously been shown to benefit from antipseudomonal therapy in general [17].

Our study findings may surprise some readers. After all, the current guidelines for the management of patients with community-acquired pneumonia recommend PCT for suspected pseudomonal pneumonia [2]. In light of new evidence, from this study, and our prior study [17], we believe the community-acquired pneumonia guidelines should be changed. We support the use of empiric PMT in high-risk patients, but we do not support the use of empiric PCT in pneumonia patients from any of the risk groups. PCT may be beneficial over PMT for patients with known Pseudomonas (i.e., definitive therapy)—a question beyond the scope of this study—but PCT is not associated with additional benefit in patients simply suspected of having Pseudomonas (i.e., empiric therapy).

Limitations

Despite the many strengths of our study, including its large sample size, risk stratification, and robust statistical methodologies to deal with dissimilar baseline characteristics, our study has limitations, most of which are inherent to all retrospective, observational studies. For instance, this study design can only identify an association, but not prove causation. It is possible that some unmeasured variable accounts for the lack of additional benefit with PCT. Given the retrospective nature of the study, we cannot ascertain the providers’ rationale for their prescribing practices. The study only included patients who received antibiotics within 48 hours of admission. Since it generally takes longer than 48 hours for culture and sensitivity results to return, we can presume this is empiric therapy, but we do not know that for sure. In addition, we acknowledge that several variables, which have been previously associated with greater disease severity in other pneumonia studies, were more prevalent among patients who received PCT. That is why we stratified patients into three risk groups and used multivariable regression models to mitigate baseline differences between groups. Risk stratification is a practical strategy, and warranted, given that we saw a benefit with antipseudomonal therapy in the high-risk group in our prior study. Regression modeling has been shown to adequately adjust for confounders in observational studies and provides adequate control for these confounders when estimating the effect of treatment on outcomes [29].

Conclusion

Older adults who received combination antipseudomonal therapy for community-onset pneumonia fared worse than those who received monotherapy. Combination therapy may be even more detrimental for low-risk patients than for medium- or high-risk patients. Empiric combination antipseudomonal therapy should not be routinely offered to all patients suspected of having Pseudomonas aeruginosa pneumonia.

Table 3.

Multivariable models to identify risk factors for 30-day mortality, including antipseudomonal monotherapy (PMT) or antipseudomonal combination therapy (PCT)

	Adjusted Odds ratio (95% confidence interval)
Risk score	All (n=31,027)	0.0-2.0 (Low Risk) (n=18,286)	2.5-7.0 (Medium Risk) (n=7,295)	7.5-21.5 (High Risk) (n=5,466)
PCT vs. PMT	1.54 (1.43-1.66)	1.69 (1.50-1.89)	1.30 (1.14-1.48)	1.21 (1.04-1.40)
Age (1-year increments)	0.91 (0.90-0.92)	0.89 (0.88-0.91)	0.92 (0.90-0.94)	0.92 (0.90-0.94)
Race	0.99 (0.90-1.10)	0.93 (0.80-1.08)	1.09 (0.92-1.30)	0.94 (0.75-1.16)
Hispanic ethnicity	0.97 (0.85-1.10)	1.06 (0.88-1.27)	0.80 (0.63-0.99)	1.04 (0.79-1.36)
Comorbid conditions
Myocardial infarction	1.13 (1.00-1.29)	1.12 (0.89-1.40)	1.02 (0.84-1.23)	0.99 (0.76-1.29)
Heart failure	1.05 (0.97-1.14)	0.98 (0.87-1.12)	0.99 (0.87-1.14)	1.00 (0.85-1.18)
COPD	0.95 (0.88-1.03)	0.95 (0.84-1.06)	0.92 (0.80-1.06)	0.92 (0.78-1.09)
Liver disease	1.32 (0.97-1.80)	0.88 (0.46-1.54)	1.20 (0.72-1.92)	1.86 (1.03-3.37)
Renal disease	0.94 (0.85-1.05)	0.89 (0.75-1.05)	0.90 (0.76-1.06)	1.01 (0.82-1.25)
Diabetes	1.03 (0.93-1.14)	1.06 (0.91-1.23)	0.98 (0.82-1.17)	1.04 (0.84-1.27)
Neoplastic disease	1.64 (1.52-1.76)	1.50 (1.35-1.67)	1.79 (1.58-2.03)	1.42 (1.21-1.67)
Medication use, by class
Cardiovascular medications	0.74 (0.68-0.79)	0.68 (0.61-0.76)	0.71 (0.61-0.81)	0.88 (0.75-1.03)
Anti-diabetic medications	0.90 (0.80-1.01)	0.78 (0.65-0.93)	1.03 (0.84-1.26)	0.96 (0.76-1.22)
Inhaled corticosteroids	0.71 (0.65-0.79)	0.67 (0.57-0.78)	0.69 (0.58-0.82)	0.88 (0.72-1.07)
Systemic corticosteroids	1.11 (1.02-1.20)	1.07 (0.94-1.23)	1.05 (0.91-1.22)	1.00 (0.83-1.19)
Pulmonary medications	0.99 (0.91-1.09)	0.99 (0.87-1.14)	1.00 (0.85-1.17)	0.91 (0.76-1.09)
Vasopressors	1.66 (1.45-1.91)	2.37 (1.81-3.09)	2.02 (1.46-2.81)	1.30 (1.08-1.57)
Mechanical ventilation
Invasive	0.88 (0.77-1.01)	2.71 (2.01-3.66)	0.85 (0.44-1.58)	0.73 (0.62-1.34)
Non-invasive	1.47 (1.28-1.70)	2.26 (1.73-2.91)	1.62 (1.13-2.29)	1.11 (0.92-0.86)
Organ failure
Respiratory	2.55 (2.29-2.84)	N/A	N/A	0.86 (0.59-1.25)
Cardiovascular	1.47 (1.30-1.66)	1.42 (1.15-1.72)	1.28 (1.00-1.62)	1.69 (1.39-2.07)
Neurological	1.36 (1.12-1.65)	1.30 (0.95-1.76)	1.47 (1.01-2.11)	1.24 (0.87-1.74)
Renal	1.58 (1.45-1.73)	1.88 (1.65-2.15)	1.42 (1.20-1.68)	1.33 (1.12-1.57)
Hematologic	1.58 (1.34-1.86)	1.44 (1.11-1.86)	1.40 (1.01-1.92)	2.00 (1.48-2.72)
Hepatic	2.55 (1.51-4.30)	2.92 (1.23-6.59)	1.97 (0.69-5.22)	2.78 (1.09-7.67)
Antibiotic therapy
GC-CAP therapy	0.53 (0.49-0.57)	0.49 (0.44-0.56)	0.59 (0.52-0.67)	0.78 (0.67-0.90)
MRSA therapy	1.22 (1.14-1.31)	1.22 (1.10-1.34)	1.19 (1.05-1.34)	1.06 (0.90-1.25)
Pseudomonas diagnosis code by discharge	0.66 (0.53-0.82)	0.71 (0.47-1.04)	0.65 (0.42-0.98)	0.65 (0.46-0.90)

Odds ratios greater than one indicate an increased risk of 30-day mortality, whereas odds ratios less than one indicate a decreased risk of 30-day mortality. Race was ordered as black versus nonblack. COPD: chronic obstructive pulmonary disease; GC-CAP: guideline-concordant community-acquired pneumonia.

Highlights.

• Retrospective cohort study of patients with community-onset pneumonia

• Antipseudomonal combination therapy has a higher mortality rate than monotherapy

• Antipseudomonal combination therapy should not be routinely offered to all patients

List of abbreviations

aOR:

adjusted odds ratio

CAP:

community-acquired pneumonia

COP:

community-onset pneumonia

CI:

confidence interval

COPD:

chronic obstructive pulmonary disease

GC:

guideline-concordant

HCAP:

healthcare-associated pneumonia

ICD-9-CM:

International Classification of Diseases, Ninth Revision, Clinical Modification codes

ICU:

intensive care unit

IV:

intravenous

IQR:

interquartile range

MDR:

multidrug-resistant

MRSA:

methicillin-resistant Staphylococcus aureus

PSE:

Pseudomonas

VHA:

Veterans Health Administration