Risk Factors for Treatment Failure in Patients Receiving Vancomycin for Hospital-Acquired Methicillin-Resistant Staphylococcus aureus Pneumonia

Abstract

Background:

The rate of vancomycin failure in patients with hospital-acquired pneumonia (HAP) caused by methicillin-resistant Staphylococcus aureus (MRSA) has exceeded 40% in several studies. This observation was attributed initially to the lack of weight-based dosing and targeting of lower trough concentrations. However, a subsequent study demonstrated no additional benefit in patients who achieved trough vancomycin concentrations >15 mg/L compared with patients with concentrations between 5 and 15 mg/L. We sought to identify contributors to vancomycin failure in patients with MRSA HAP.

Methods:

This was a retrospective study of patients in a surgical intensive care unit with MRSA HAP who received vancomycin between January 1, 2005, and July 31, 2007. Clinical outcomes, microbiological data, prior antibiotic exposure, ventilator days, co-morbidities, and demographics were compared in patients with clinical success and those with treatment failure. Their characteristics were compared using a two-sided Fisher exact test or Mann-Whitney U test, as appropriate for nominal or continuous data.

Results:

More patients in the treatment failure group had received one or more doses of vancomycin within 90 days leading up to MRSA HAP (84% vs. 47%; p = 0.04). In addition, the duration of prior vancomycin exposure was significantly longer among patients in the treatment failure group (6 vs. 0 days; p < 0.05). There were no statistically significant differences in the percentages of patients who achieved a vancomycin trough concentrations ≥15 mg/dL within the first 48 h (28% vs. 17%; p = 0.69), 72 h (44% vs. 39%; p = 1.0), or 96 h (56% vs. 44%; p = 0.74) after starting treatment. Patients in the failure group had a significantly higher overall mortality rate (32% vs. 0; p = 0.02).

Conclusions:

These data suggest that patients who have recent exposure to vancomycin are at high risk for vancomycin failure and may benefit from an appropriate alternative when a diagnosis of MRSA HAP is made.

Hospital-acquired pneumonia (HAP) is a significant contributor to hospital morbidity and death, with an estimated mortality rate ranging from 33% to 50% depending on the pathogen [1 –5]. Each episode of HAP prolongs hospital length of stay (LOS) and increases health care expenditure by $40,000 [1, 6]. Staphylococcus aureus is the most common pathogen in HAP, accounting for approximately 30% of all cases [1]. In addition, greater than 50% of Staphylococcus aureus isolates in the intensive care unit (ICU) setting are methicillin resistant.

Vancomycin is the standard of care for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) pneumonia; however, the rate of treatment failure has exceeded 40% in several studies [7 –9]. Initial expert opinion attributed the high rate of failure to the lack of weight-based vancomycin dosing and potential targeting of lower trough concentrations. On the basis of this assumption and the fact that vancomycin has relatively poor tissue penetration in the lung, published guidelines recommend targeting vancomycin trough concentrations between 15 and 20 mg/dL for this population [1, 10, 11]. However, prospective data supporting this recommendation are lacking. Furthermore, a subsequent retrospective analysis by Jeffres et al. demonstrated no benefit as judged by mortality rate reduction among patients with MRSA HAP who achieved trough concentrations >15 mg/L compared with patients with concentrations between 5 and 15 mg/L, suggesting that troughs may not correlate with clinical outcome [12].

Additional data published in a retrospective analysis of two randomized controlled trials have demonstrated a higher rate of treatment failure among patients with MRSA pneumonia who have multiple lobe involvement, ventilator-associated pneumonia (VAP), renal dysfunction, or oncologic co-morbidities [13]. A similar study demonstrated that hepatic dysfunction, vascular co-morbidities, multiple lobe pneumonia, and Acute Physiology and Chronic Health Evaluation (APACHE II) score >20 were independent risk factors for treatment failure in MRSA VAP [14]. However, these studies included patients who were treated with vancomycin or linezolid and did not address risk factors for treatment failure specific to vancomycin.

Given the lack of understanding of the high rate of clinical failure associated with vancomycin therapy for MRSA HAP, the objective of our study was to identify independent risk factors and treatment characteristics that correlate with treatment failure in these patients. These risk factors may assist in predicting which patients will benefit from initial treatment with alternatives to vancomycin. We also designed the study to assess the effect of vancomycin failure on clinical outcomes.

Patients and Methods

Vanderbilt University Hospital (VUH) in Nashville, TN, is an integrated academic tertiary-care medical center with a 21-bed adult surgical ICU (SICU) that admits approximately 1,300 patients annually. The mean APACHE II severity of illness score on admission to the SICU is 17, and the average LOS is 3.5 days. We conducted a retrospective investigation of vancomycin treatment for MRSA HAP in patients admitted to the SICU between January 1, 2005, and July 31, 2007. Trauma patients, who are admitted to an ICU maintained independently from the SICU, were not enrolled. The study was approved by the Institutional Review Board at VUH.

As a standard practice in our SICU, bronchoscopy is performed in ventilated patients with suspected pneumonia, during which bronchoalveolar lavage (BAL) samples are collected for quantitative culture and susceptibility testing. Immediately after BAL sample collection, broad-spectrum therapy is initiated in accordance with nationally published guidelines utilizing computer order entry order sets designed to cover likely pathogens empirically on the basis of ICU-specific historical culture and susceptibility data [1]. Unless contraindicated, until culture results become available, all patients with presumed HAP in the SICU receive initial treatment with three antibiotics: Vancomycin in combination with an anti-pseudomonal fluoroquinolone or an aminoglycoside plus either an anti-pseudomonal β-lactam/β-lactamase inhibitor combination, anti-pseudomonal cephalosporin, or anti-pseudomonal carbapenem. Therapy is discontinued, narrowed, or tailored once culture and susceptibility results are finalized. By protocol, patients who do not improve clinically after 96 h of treatment, as determined by radiographic findings, ventilator requirements, the presence of fever, leukocytosis, sputum characteristics, or the presence of rales routinely undergo repeat bronchoscopy with BAL sample collection.

Prior to the study period, a validated vancomycin empiric dosing nomogram was implemented in the computerized physician order entry system at VUH. The nomogram provides dosing recommendations on the basis of each patient's weight and renal function. For example, patients with normal renal function (i.e., estimated creatinine clearance >110 mL/min) receive initial total daily doses between 25 and 50 mg/kg divided into two or three doses depending on the patient's weight. Patients receiving intermittent or continuous renal replacement therapy and individuals <40 kg or >120 kg are excluded from the nomogram and receive vancomycin as determined by the SICU clinical pharmacist. In most patients with HAP, vancomycin trough concentrations are measured before the fourth or fifth dose, and doses are adjusted as needed to target a serum trough between 15 and 20 mg/L. Patients requiring dialysis receive intermittent doses of vancomycin when the serum concentration falls below 20 mg/L. Vancomycin is discontinued after eight days in patients with documented MRSA pneumonia who respond adequately to treatment, and vancomycin is converted to linezolid at the discretion of the attending physician in patients failing vancomycin.

Patients were identified for study enrollment utilizing an SICU informatics-driven database that captures information on all patients admitted to the unit, including outcome variables, severity scoring variables, and all infection data reported to the National Nosocomial Infections Surveillance System/National Healthcare Safety Network (NNIS/NHSN) [15]. Dedicated infection control practitioners within each ICU at VUH maintain continuous infection surveillance. Patients were included in the study if they met the PNU1 (Table 1) or PNU2 (Table 2) NNIS/NHSN criteria for HAP, received treatment with vancomycin for ≥96 h, and were admitted to the SICU between January 1, 2005, and July 31, 2007. Exclusion criteria were age <18 years or pneumonia resulting from a pathogen other than MRSA. A total of 217 patients admitted to the SICU during the enrollment period met the NNIS/NHSN criteria for HAP. Staphylococcus aureus was the pathogen in 29% of these cases (n = 63), and MRSA accounted for 61% (n = 39) of Staphylococcus aureus isolates. One patient with MRSA HAP never received vancomycin and was excluded from the study. One patient developed MRSA HAP during two separate SICU admissions >12 months apart, and each episode was assessed independently and included in the analysis.

Table 1.

Specific Site Algorithms for Clinically Defined Pneumonia (PNU1) ^a

Radiology findings	Signs/symptoms/laboratory
Two or more serial chest radiographs with at least one of the following^b,c,d New or progressive and persistent infiltrate Consolidation Cavitation	FOR ANY PATIENT, at least one of the following: • Fever (>38°C or >100.4°F) with no other recognized cause • Leukopenia (<4,000 white blood cells [WBC]/mm³) or leukocytosis (≥12,000 WBC/mm³) • For adults ≥70 years old, altered mental status with no other recognized cause
	And at least two of the following: • New onset of purulent sputum,^e or change in character of sputum,^f or increased suctioning requirements • New onset or worsening cough, or dyspnea, or tachypnea^g • Rales^h or bronchial breath sounds • Worsening gas exchange (e.g., O₂ desaturation (e.g., PaO₂/FiO₂ ≤240),ⁱ increased oxygen requirements, or increased ventilator demand)

Adapted from a publication of the Centers for Disease Control and Prevention [15].

Occasionally, in non-ventilated patients, the diagnosis of nosocomial pneumonia is clear on the basis of symptoms, signs, and a single definitive chest radiograph. However, in patients with pulmonary or cardiac disease (for example, interstitial lung disease or congestive heart failure), the diagnosis of pneumonia may be particularly difficult. Other non-infective conditions (for example, pulmonary edema from decompensated congestive heart failure) may simulate the presentation of pneumonia. In these more difficult cases, serial chest radiographs must be examined to help separate infective from non-infective pulmonary processes. To help confirm difficult cases, it may be useful to review radiographs on the day of diagnosis, three days prior to the diagnosis, and on days 2 and 7 after the diagnosis. Pneumonia may have a rapid onset and progression, but does not resolve quickly. Radiographic changes persist for several weeks. As a result, rapid radiographic resolution suggests that the patient does not have pneumonia, but rather a non-infective process such as atelectasis or congestive heart failure.

Note that there are many ways of describing the radiographic appearance of pneumonia. Examples include, but are not limited to, “air-space disease,” “focal opacification,” and “patchy areas of increased density.” Although perhaps not specifically delineated as pneumonia by the radiologist, in the appropriate clinical setting, these alternative descriptive wordings should be considered seriously as potential descriptions of pneumonia.

In patients without underlying pulmonary or cardiac disease (e.g., respiratory distress syndrome, bronchopulmonary dysplasia, pulmonary edema, or chronic obstructive pulmonary disease), one definitive chest radiograph is acceptable.

Purulent sputum is defined as secretions from the lungs, bronchi, or trachea that contain >25 neutrophils and <10 squamous epithelial cells per low-power field (×100). If the laboratory reports these data qualitatively (e.g., “many WBCs” or “few squames”), be sure their descriptors match this definition of purulent sputum. This laboratory confirmation is required because written clinical descriptions of purulence are highly variable.

A single notation of either purulent sputum or change in the character of the sputum is not meaningful; repeated notations over a 24-h period would be more indicative of the onset of an infectious process. “Change in the character of sputum” refers to the color, consistency, odor, and quantity.

In adults, tachypnea is defined as a rate >25 breaths/min. Tachypnea is defined as >75 breaths/min in infants born at <37 weeks' gestation and until the 40th week; >60 breaths/min in patients <2 months old; >50 breaths/min in patients 2–12 months old; and >30 breaths/min in children >1 year old.

Rales may be described as “crackles.”

This measure of arterial oxygenation is defined as the ratio of the arterial tension (PaO₂) to the inspiratory fraction of oxygen (FiO₂).

Notes: Care must be taken to determine the etiology of pneumonia in a patient with positive blood cultures and radiographic evidence of pneumonia, especially if the patient has invasive devices such as intravascular lines or an indwelling urinary catheter. In general, in an immunocompetent patient, coagulase-negative staphylococci, common skin contaminants, and yeasts in blood cultures will not be the etiologic agent of the pneumonia. An endotracheal aspirate is not a minimally contaminated specimen. Therefore, an endotracheal aspirate does not meet the laboratory criteria.

Table 2.

Specific Site Algorithms for Pneumonia with Common Bacterial or Filamentous Fungal Pathogens and Specific Laboratory Findings (PNU2)^a

Radiologic signs	Signs/symptoms	Laboratory findings
Two or more serial chest radiographs with at least one of the following^b,c,d New or progressive and persistent infiltrate Consolidation Cavitation	At least one of the following: Fever (>38°C or >100.4°F) with no other recognized cause Leukopenia (<4,000 white blood cells [WBC]/mm³) or leukocytosis (≥12,000 WBC/mm³) For adults ≥70 years old, altered mental status with no other recognized cause And at least one of the following: New onset of purulent sputum^e or change in character of sputum^f or increased respiratory secretions or greater suctioning requirements New onset or worsening cough, dyspnea, or tachypnea^g Rales^h or bronchial breath sounds Worsening gas exchange (e.g., O₂ desaturation (e.g., PaO₂/FiO₂ ≤240),ⁱ increased oxygen requirements, or increased ventilator demand)	At least one of the following: Growth in blood culture^j not related to another source of infection Growth in culture of pleural fluid Positive quantitative culture^k from minimally contaminated LRT specimen (e.g., BAL or protected specimen brushing) ≥5% BAL-obtained cells contain bacteria on direct microscopic examination (e.g., Gram stain)Histopathologic examination shows at least one of the following signs of pneumonia: Abscess formation or foci of consolidation with intense PMN accumulation in bronchioles and alveoli Positive quantitative culture^k of lung parenchyma Evidence of lung parenchyma invasion by fungal hyphae or pseudohyphae

Adapted from a publication of the Centers for Disease Control and Prevention [15].

Occasionally, in non-ventilated patients, the diagnosis of nosocomial pneumonia is clear on the basis of symptoms, signs, and a single definitive chest radiograph. However, in patients with pulmonary or cardiac disease (for example, interstitial lung disease or congestive heart failure), the diagnosis of pneumonia may be particularly difficult. Other non-infective conditions (for example, pulmonary edema from decompensated congestive heart failure) may simulate the presentation of pneumonia. In these more difficult cases, serial chest radiographs must be examined to help separate infective from non-infective pulmonary processes. To help confirm difficult cases, it may be useful to review radiographs on the day of diagnosis, three days prior to the diagnosis, and on days 2 and 7 after the diagnosis. Pneumonia may have a rapid onset and progression but does not resolve quickly. Radiographic changes of pneumonia persist for several weeks. As a result, rapid radiographic resolution suggests that the patient does not have pneumonia, but rather a non-infective process such as atelectasis or congestive heart failure.

Note that there are many ways of describing the radiographic appearance of pneumonia. Examples include, but are not limited to, “air-space disease,” “focal opacification,” and “patchy areas of increased density.” Although perhaps not specifically delineated as pneumonia by the radiologist, in the appropriate clinical setting these alternative descriptive wordings should be considered seriously as potential descriptions of pneumonia.

A single notation of either purulent sputum or a change in the character of the sputum is not meaningful; repeated notations over a 24-h period would be more indicative of the onset of an infective process. “Change in the character of sputum” refers to the color, consistency, odor and quantity.

In adults, tachypnea is defined as a rate of >25 breaths/min. Tachypnea is defined as >75 breaths/min in infants born at <37 weeks' gestation and until the 40th week; >60 breaths/min in patients <2 months old; >50 breaths/min in patients 2–12 months old; and >30 breaths/min in children >1 year old.

Rales may be described as “crackles.”

This measure of arterial oxygenation is defined as the ratio of the arterial tension (PaO₂) to the inspiratory fraction of oxygen (FiO₂).

Care must be taken to determine the etiology of pneumonia in a patient with positive blood cultures and radiographic evidence of pneumonia, especially if the patient has invasive devices such as intravascular lines or an indwelling urinary catheter. In general, in an immunocompetent patient, coagulase-negative staphylococci, common skin contaminants, and yeasts in blood cultures will not be the etiologic agent of the pneumonia.

An endotracheal aspirate is not a minimally contaminated specimen. Therefore, an endotracheal aspirate does not meet the laboratory criteria.

LRT = lower respiratory tract; BAL = bronchoalveolar lavage; PMN = polymorphonuclear leukocytes.

Information regarding cultures and susceptibilities, infection history, prior antibiotic exposure, vancomycin dosing and trough characteristics, requirement for renal replacement therapy, ventilator and hospital days prior to infection, APACHE II scores, co-morbidities, and demographic data were collected for each patient. We also assessed the following outcomes data: patient disposition, hospital and ICU LOS, ventilator-free days, requirement for vasopressor support, and conversion from vancomycin to linezolid. Ventilator-free days were calculated using a 28-day landmark as described by Schoenfeld et al. [16].

Patients enrolled in the study were assessed for clinical response to vancomycin therapy until death or hospital discharge. Characteristics were compared in the group of patients with clinical success and those who failed treatment. Vancomycin failure was defined as persistence, progression, or recurrence of MRSA pneumonia or infection-related death. Patients were considered to have persistence, progression, or recurrence of pneumonia when signs of clinical improvement were not achieved or maintained and a BAL sample obtained ≥96 h after initiation of vancomycin therapy grew ≥10⁴ MRSA colony-forming units (CFU)/mL. In patients with a fatal outcome, the cause of death was determined during an independent review by a physician unaware of the treatment given.

Data were expressed as numbers (percentage) or medians (interquartile range; [IQR]). Characteristics were compared using a two-sided Fisher exact test or Mann-Whitney U test as appropriate for nominal or continuous data, with an alpha of 0.05. All statistical analyses were performed with SPSS software (version 14.0 for Windows, SPSS Inc., Chicago, IL).

Results

Outcomes

The baseline characteristics of the study population are presented in Table 3. In the 38 patients meeting the inclusion criteria, the rate of vancomycin treatment failure was 50%. The most common reason for failure was persistence, progression, or recurrence of MRSA HAP (n = 14). Five patients suffered infection-related death; an additional death was unrelated to the patient's pneumonia. Within the failure group, nine patients developed breakthrough MRSA HAP, which was defined as an initial BAL culture positive for MRSA in patients who had received vancomycin for ≥96 h. In addition, eight patients developed MRSA HAP again after completing a course of vancomycin therapy.

Table 3.

Demographic and Clinical Characteristics and Microbiological Data for Patients with Clinical Cure or Treatment Failure ^a

Characteristic	Clinical cure (n = 19)	Treatment failure (n = 19)
Median age (years) (IQR)	64 (47–71)	62 (51–76)
Female sex (%)	8 (42)	3 (16)
Median weight (kg) (IQR)	78 (74–94)	80 (54–103)
Median body mass index (kg/m²) (IQR)	26 (25–35)	27 (20–32)
Median APACHE II score (IQR)	15 (13–20)	18 (16–22)
Cystic fibrosis (%)	0	1 (5)
Lung transplant (%)	0	2 (11)
VRE infection (any site) (%)	0	1 (5)
MRSA bacteremia (%)	3 (16)	2 (11)
Initial MRSA isolate susceptible to (%):^b
Clindamycin	0	2 (11)
Erythromycin	0	0
Gentamicin	16 (84)	17 (90)
Levofloxacin	3 (16)	0
TMP-SMX	19 (100)	19 (100)
Vancomycin	19 (100)	19 (100)

None of the differences are statistically significant.

Determined by the Kirby-Bauer method.

APACHE II = Acute Physiology and Chronic Health Evaluation; IQR = interquartile range; MRSA = methicillin-resistant Staphylococcus aureus; TMP-SMX = trimethoprim/sulfamethoxazole; VRE = vancomycin-resistant Enterococcus.

There were no statistically significant differences between the treatment failure and clinical cure groups. One patient in the clinical failure group had cystic fibrosis and was admitted to the SICU after lung transplantation, and one additional patient in this group developed MRSA HAP in a lung allograft.

Contributors to vancomycin failure

A comparison of potential risk factors for vancomycin failure is shown in Table 4. In regard to antibiotic exposure prior to MRSA HAP, the results in the two groups were similar except that more patients in the treatment failure group had received one or more doses of vancomycin within the 90 days leading up to MRSA HAP (84% vs. 47%; p = 0.04). In addition, the duration of prior vancomycin exposure was significantly longer among patients in the treatment failure group (6 vs. 0 days; p < 0.01), and more patients who failed therapy had received previously vancomycin treatment for ≥96 h (58% vs. 5%; p = 0.001).

Table 4.

Comparison of Risk Factors in Patients with Clinical Cure or Vancomycin Failure

Factor	Clinical cure (n = 19)	Treatment failure (n = 19)	P value
Hospitalization within prior 90 days (%)	9 (47)	6 (32)	0.51
Prior MRSA infection (any site) (%)	0	3 (16)	0.23
Any renal replacement therapy (%)	2 (11)	4 (21)	0.66
Intermittent dialysis	1 (5)	1 (5)	1.0
Continuous renal replacement	2 (11)	3 (16)	1.0
Median hospital days prior to infection (IQR)	7 (5–11)	12 (7–20)	0.11
Median ICU days prior to infection (IQR)	5 (2–10)	7 (4–11)	0.26
Mechanical ventilation (%)	19 (100)	19 (100)	1.0
Mechanical ventilation ≥48 h (%)	17 (89)	19 (100)	0.49
Median ventilator days prior to infection (IQR)	4 (3–8)	7 (5–10)	0.06
Previous antibiotic exposure (%)^a	17 (90)	16 (84)	1.0
Vancomycin	9 (47)	16 (84)	0.04
Penicillin	4 (21)	5 (26)	1.0
Cephalosporin	10 (53)	7 (37)	0.52
Carbapenem	3 (16)	3 (16)	1.0
Aminoglycoside	2 (11)	4 (21)	0.66
Fluoroquinolone	9 (47)	12 (63)	0.52
Macrolide	1 (5)	4 (21)	0.34

Within 90 days prior to infection.

ICU = intensive care unit; IQR = interquartile range; MRSA = methicillin-resistant Staphylococcus aureus.

We also compared vancomycin dosing and trough characteristics. The starting dose was similar in the treatment failure and clinical success groups (20 vs. 25 mg/kg/day; p = 0.23), and there was no difference in the duration of vancomycin therapy (9 vs. 8 days; p = 0.66). The initial trough concentration, which most frequently was measured prior to the fourth or fifth dose, was similar in the two groups (13 vs. 13 mg/dL; p = 0.91), as was the mean trough over the duration of vancomycin therapy (16 vs. 16 mg/dL; p = 0.57). We also assessed the time to reach a vancomycin trough ≥15 mg/dL. There were no statistically significant differences in the percentages of patients who achieved a therapeutic trough within the first 48 h (28% vs. 17%; p = 0.69), 72 h (44% vs. 39%; p = 1.0), or 96 h (56% vs. 44%; p = 0.74) after starting vancomycin treatment. Additionally, a similar proportion of patients in the treatment failure and clinical success groups achieved a therapeutic vancomycin trough concentration at least once during the treatment course (67% vs. 78%; p = 0.71).

All patients in our study population were intubated during their hospitalization, and a similar percentage of patients in both groups required mechanical ventilation for ≥48 h (100 vs. 89%; p = 0.49). Patients in our study were most commonly intubated perioperatively before transfer to the SICU. Patients in the treatment failure group had more hospital days (12 vs. 7; p = 0.11), ICU days (7 vs. 5; p = 0.26), and ventilator days (7 vs. 4; p = 0.06) prior to the initial MRSA-positive BAL culture; however, these differences did not reach statistical significance. There also was a non-significant trend toward a higher percentage of patients in the vancomycin failure group having a previous documented MRSA infection (16% vs. 0; p = 0.23).

Effect of vancomycin failure on clinical outcomes

Clinical outcomes in patients with clinical success and patients with vancomycin failure are compared in Table 5. Patients in the failure group had a significantly higher number of vasopressor days (10 vs. 1; p = 0.02) and overall mortality rate (32% vs. 0; p = 0.02). In addition, there were significantly fewer ventilator-free days among the patients in the failure group than in the clinical success group (0 vs. 17 days; p = 0.04). There was a trend toward a higher rate of ICU death (21% vs. 0) among the treatment failure group, although this difference did not reach statistical significance (p = 0.11). We did not identify any differences in hospital LOS (35 vs. 24 days; p = 0.53) or ICU LOS (15 vs. 20 days; p = 0.26) when non-survivors were removed from the analysis.

Table 5.

Clinical Outcomes in Patients with Clinical Cure or with Vancomycin Failure

Outcome	Clinical cure (n = 19)	Treatment failure (n = 19)	P value
Death (%)	0	6 (32)	0.02
ICU death (%)	0	4 (21)	0.11
Median hospital stay (days) (IQR)	24 (20–39)	27 (18–46)	0.99
Median ICU stay (days) (IQR)	20 (12–31)	15 (9–25)	0.25
Median ventilator days after infection onset (IQR)	6 (3–12)	6 (3–22)	0.74
Median total ventilator days (IQR)	11 (9–19)	17 (10–27)	0.46
Median ventilator-free days (IQR)	17 (10–19)	0 (0–16)	0.04
Vasopressor support (%)	15 (79)	19 (100)	0.11
Median vasopressor days (IQR)	1 (0–10.5)	10 (6–24)	0.02
Conversion to linezolid (%)	8 (42)	8 (42)	1.0

ICU = intensive care unit; IQR = interquartile range.

The rate of conversion from vancomycin to linezolid was identical in the treatment failure and clinical success groups (42%). One patient in the clinical success group was converted to oral linezolid after eight days of vancomycin because of the loss of intravenous access. Another patient in the clinical success group was converted to oral linezolid on the day of hospital discharge to avoid the need to administer antibiotics intravenously at home. This patient received vancomycin for five days prior to conversion to linezolid.

Discussion

Published data examining risk factors for treatment failure in patients with MRSA pneumonia are limited despite the high rate of vancomycin failure reported in the literature. Our study was designed to describe the potential risk factors and treatment characteristics associated with vancomycin failure in this population. We found that prior exposure to the antibiotic increases the risk of vancomycin treatment failure significantly, with the percentage of patients who received one or more doses of vancomycin within 90 days before developing pneumonia being significantly higher in the treatment failure group. The extent of previous exposure also appears to be an important determining factor for treatment outcome, as patients who failed therapy were significantly more likely to have received ≥96 h of vancomycin prior to developing MRSA HAP.

To our knowledge, an association between prior vancomycin exposure and treatment failure has not been reported, and the mechanism contributing to such an association is not clear. The most plausible explanation is that MRSA organisms that survive initial exposure to vancomycin develop mechanisms to increase the minimum inhibitory concentration (MIC), such as thickening of the peptidoglycan cell wall [9]. Other factors associated with higher vancomycin MICs among MRSA isolates include an upward drift over time and a concomitant reduction in clindamycin, doxycycline, erythromycin, or rifampin susceptibilities [17 –20].

It is possible that higher MICs reduce the effectiveness of vancomycin, even when the MIC remains below the resistance breakpoint of 8 mcg/mL [21, 22]. Indeed, Hidayat et al. demonstrated a significantly lower end-of-treatment response and a trend toward an increase in the infection-related mortality rate among patients with an MRSA isolate with an MIC ≥2 mcg/mL compared with strains with lower MICs [23]. Additional data suggest the rate of treatment success with vancomycin falls to <10% when the MIC of the MRSA isolate reaches 1 mcg/mL or greater [24]. Furthermore, patients with bacteremia are 2.4 times more likely to experience vancomycin failure when the MRSA MIC is ≥1.5 mcg/mL [25].

All of the MRSA isolates in our study population were determined to be susceptible to vancomycin, with an MIC ≤8 mcg/mL. However, one limitation to our study is that all susceptibility testing was performed exclusively by the Kirby-Bauer method, which unfortunately does not provide quantitative MIC data and is unable to differentiate strains with reduced susceptibility to vancomycin (MIC range 4–8 mcg/mL) from fully susceptible strains (MIC ≤2 mcg/mL) [18, 21]. Thus, our hypothesis that previous vancomycin exposure leads to higher MICs requires further investigation. Our results, in combination with published literature, have prompted our institution to begin routine collection of MIC data on all MRSA isolates. Future studies are needed to assess the correlation of vancomycin exposure and MIC values among MRSA isolates obtained from BAL cultures.

Although inadequate dosing likely would lead to treatment failure, we did not identify an association between treatment response and vancomycin dosing or trough values in our population. Suboptimal vancomycin dosing, represented by the pharmacodynamic ratio of the area under the curve to the MIC (AUC/MIC) ≤345, has been shown to decrease the rate of clinical response in patients with lower respiratory tract MRSA infections [26]. Assuming no differences in pharmacokinetic parameters between the treatment failure and clinical success groups in our study, the mean AUC likely was similar in the two groups, as the patients received similar vancomycin doses (mg/kg/day).

Although the AUC/MIC ratio is the best pharmacodynamic parameter to describe the activity of vancomycin, monitoring serum trough concentrations is the clinical standard of care. Despite expert recommendations to target a vancomycin trough between 15 and 20 mg/dL for patients with MRSA pneumonia, prospective data demonstrating a benefit of higher trough concentrations are limited [1, 10]. In addition, a recent retrospective analysis by Jeffres et al. demonstrated absence of an association between higher trough concentrations and a reduced mortality rate [12]. One limitation of their study was that the time to achieve a trough >15 mg/L was not assessed. Thus, we were compelled to evaluate all aspects of vancomycin trough concentrations. Similar to Jeffres et al., we did not find an association between the percentage of patients who achieved a trough ≥15 mg/dL and treatment outcomes. We also found no correlation between the initial trough or time to achieve a trough ≥15 mg/dL in the treatment failure and clinical success groups. Thus, additional data are needed to identify any relation between vancomycin trough characteristics and the probability of treatment success in patients with MRSA pneumonia.

Unlike previous studies describing a relation between renal insufficiency and vancomycin failure secondary to underdosing because of fear of drug accumulation, we found no such association [10, 13, 26]. The absence of such a correlation in our study likely is related to several factors. First, we assessed only renal failure that necessitated intermittent hemodialysis or continuous renal replacement therapy. We did not compare creatinine clearances in the two groups. Second, much of the apprehension about dosing vancomycin in patients with renal failure was removed from our SICU by the implementation of a validated dosing nomogram.

A potential limitation of this study is our reliance on the medical record for data collection, which is dependent on the accuracy and thoroughness of documentation for appropriate data retrieval. However, the difficulty in evaluating each medical chart retrospectively was minimized by our access to the SICU repository and electronic medical record at VUH. Another limitation is the small sample size, which reduces our power to detect potentially relevant differences in confounders. However, despite the relatively small population, this study was powered adequately to identify certain risk factors for treatment failure, such as previous vancomycin exposure, as the difference between groups reached statistical significance. One other potential limitation of this analysis is our definition of treatment failure, which tends to differ among papers. Several other studies in this population have incorporated administration of a non-study antibiotic for MRSA pneumonia into the criteria for vancomycin failure. We specifically chose not to categorize patients who received linezolid in the treatment failure group automatically, as several patients were converted from vancomycin for reasons other than treatment failure, and because this enabled us to assess the prescribing patterns of clinicians at our institution better.

Conclusions

Although there are published data assessing the correlation between vancomycin troughs and clinical outcomes, our study is among the first to report the absence of an association between treatment response and the time to achieve a vancomycin trough ≥15 mg/dL. In addition, our results demonstrate that vancomycin exposure within 90 days leading up to infection has the greatest propensity to elicit treatment failure in patients with MRSA HAP receiving the drug. The duration and extent of prior vancomycin exposure also correlated with the rate of treatment success, and patients who do not respond to treatment have significantly worse clinical outcomes. Our data suggest that patients who have recent exposure to vancomycin may benefit from an appropriate alternative when a diagnosis of MRSA HAP is made. Further studies with a prospective design may provide additional important information regarding risk factors for vancomycin failure.

Footnotes

Author Disclosure Statement

C.B.C. has served as an Advisory Board Member for Novartis and Pfizer and has received grant support from Pfizer, Merck, Cubist, and Astra Zeneca. No other competing financial interests exist.

References

Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med, 2005; 171:388–416.

Shorr

, Combes

, Kollef

, Chastre

. Methicillin-resistant Staphylococcus aureus prolongs intensive care unit stay in ventilator-associated pneumonia, despite initially appropriate antibiotic therapy. Crit Care Med, 2006; 34:700–706.

Fagon

, Chastre

, Hance

et al. Nosocomial pneumonia in ventilated patients: A cohort study evaluating attributable mortality and hospital stay. Am J Med, 1993; 94:281–288.

Heyland

, Cook

, Griffith

et al. The attributable morbidity and mortality of ventilator-associated pneumonia in the critically ill patient. The Canadian Critical Trials Group. Am J Respir Crit Care Med, 1999; 159:1249–1256.

Tablan

, Anderson

, Besser

et al. Guidelines for preventing health-care-associated pneumonia, 2003: Recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee. MMWR Recomm Rep, 2004; 53:1–36.

Rello

, Ollendorf

, Oster

et al. Epidemiology and outcomes of ventilator-associated pneumonia in a large US database. Chest, 2002; 122:2115–2121.

Fagon

, Patrick

, Haas

et al. Treatment of gram-positive nosocomial pneumonia: Prospective randomized comparison of quinupristin/dalfopristin versus vancomycin. Nosocomial Pneumonia Group. Am J Respir Crit Care Med, 2000; 161:753–762.

Ioanas

, Lode

. Linezolid in VAP by MRSA: A better choice? Intensive Care Med, 2004; 30:343–346.

Lam

, Wunderink

. Methicillin-resistant S. aureus ventilator-associated pneumonia: Strategies to prevent and treat. Semin Respir Crit Care Med, 2006; 27:92–103.

10.

Craven

, Palladino

, McQuillen

. Healthcare-associated pneumonia in adults: Management principles to improve outcomes. Infect Dis Clin North Am, 2004; 18:939–962.

11.

Scheetz

, Wunderink

et al. Potential impact of vancomycin pulmonary distribution on treatment outcomes in patients with methicillin-resistant Staphylococcus aureus pneumonia. Pharmacotherapy, 2006; 26:539–550.

12.

Jeffres

, Isakow

, Doherty

et al. Predictors of mortality for methicillin-resistant Staphylococcus aureus health-care-associated pneumonia: Specific evaluation of vancomycin pharmacokinetic indices. Chest, 2006; 130:947–955.

13.

Wunderink

, Rello

, Cammarata

et al. Linezolid vs vancomycin: Analysis of two double-blind studies of patients with methicillin-resistant Staphylococcus aureus nosocomial pneumonia. Chest, 2003; 124:1789–1797.

14.

Kollef

, Rello

, Cammarata

et al. Clinical cure and survival in gram-positive ventilator-associated pneumonia: Retrospective analysis of two double-blind studies comparing linezolid with vancomycin. Intensive Care Med, 2004; 30:388–394.

15.

Division of Healthcare Quality Promotion; National Center for Infectious Diseases. The National Healthcare Safety Network (NHSN) Manual: Patient Safety Component Protocol, 2008; 1–98.

16.

Schoenfeld

, Bernard

. Statistical evaluation of ventilator-free days as an efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med, 2002; 30:1772–1777.

17.

Wang

, Hindler

, Ward

, Bruckner

. Increased vancomycin MICs for Staphylococcus aureus clinical isolates from a university hospital during a 5-year period. J Clin Microbiol, 2006; 44:3883–3886.

18.

Mohr

, Murray

. Point: Vancomycin is not obsolete for the treatment of infection caused by methicillin-resistant Staphylococcus aureus. Clin Infect Dis, 2007; 44:1536–1542.

19.

Steinkraus

, White

, Friedrich

. Vancomycin MIC creep in non-vancomycin-intermediate Staphylococcus aureus (VISA), vancomycin-susceptible clinical methicillin-resistant S. aureus (MRSA) blood isolates from 2001–05. J Antimicrob Chemother, 2007; 60:788–794.

20.

Deresinski

. Counterpoint: Vancomycin and Staphylococcus aureus—An antibiotic enters obsolescence. Clin Infect Dis, 2007; 44:1543–1548.

21.

Performance standards for antimicrobial susceptibility testing; seventeenth informational supplement. Clinical Laboratory Standards Institute DocumentM100-S17 [ISBN 1-56238-625-5]2007; 27:1–182.

22.

Cosgrove

, Carroll

, Perl

. Staphylococcus aureus with reduced susceptibility to vancomycin. Clin Infect Dis, 2004; 39:539–545.

23.

Hidayat

, Hsu

, Quist

et al. High-dose vancomycin therapy for methicillin-resistant Staphylococcus aureus infections: Efficacy and toxicity. Arch Intern Med, 2006; 166:2138–2144.

24.

Sakoulas

, Moise-Broder

, Schentag

et al. Relationship of MIC and bactericidal activity to efficacy of vancomycin for treatment of methicillin-resistant Staphylococcus aureus bacteremia. J Clin Microbiol, 2004; 42:2398–2402.

25.

Lodise

, Graves

, Evans

et al. Relationship between vancomycin MIC and failure among patients with methicillin-resistant Staphylococcus aureus bacteremia treated with vancomycin. Antimicrob Agents Chemother, 2008; 52:3315–3320.

26.

Moise

, Forrest

, Bhavnani

et al. Area under the inhibitory curve and a pneumonia scoring system for predicting outcomes of vancomycin therapy for respiratory infections by Staphylococcus aureus. Am J Health Syst Pharm, 2000; 57:S4–S9.