Preventing Ventilator-Associated Pneumonia in Adults^*: Sowing Seeds of Change FREE TO VIEW

The best time to plant a tree is 20 years ago. The second best time is now.

African Proverb

Ventilator-associated pneumonia (VAP) is common and associated with high morbidity, mortality, and health-care costs, estimated to be $40,000/case.¹²³ VAP has a cumulative incidence of 10 to 25% and accounts for approximately 25% of all ICU infections and > 50% of the antibiotics prescribed in ICU, making it a primary focus for risk-reduction strategies. Crude mortality rates for VAP may be as high as 20 to 70% and are generally highest in medical ICU patients and those with bacteremia or pneumonia due to multidrug-resistant (MDR) pathogens, such as Pseudomonas aeruginosa or Acinetobacter species, and methicillin-resistant Staphylococcus aureus (MRSA).¹²

Due to the limitations on space and the number of citations for this review, highlights from selected guidelines and publications on VAP prevention in adults, primarily published after January 1, 2003, were included.¹²⁴ Emphasis was placed on new concepts, controversies, and barriers to implementing beneficial, cost-effective programs aimed at reducing VAP and improving patient outcomes.

As with the general population, hospitalized patients are now older and have more comorbidities, immune dysfunction, invasive procedures, and exposure to antibiotics. Patients are also increasingly mobile and more likely to reside in short-term and long-term health-care settings, increasing the risk of colonization, person-to-person transmission, and infection with MDR pathogens.⁵ MDR Gram-negative bacilli, such as P aeruginosa, Klebsiella pneumoniae, and Acinetobacter species, often result into higher patient mortality, longer hospital stays, poorer functional status at discharge, and need for transitional care⁶⁷

MRSA now accounts for > 50% of the ICU-acquired staphylococcal infections in the United States, is associated with significant mortality and morbidity, and is a challenge for infection control teams.⁸ The rapid emergence of community-acquired MRSA infection and its increasing role in hospitals is of great concern, as these isolates are genetically distinct from hospital strains, and often carry the Pantin-Valentine leukocidin virulence factor, which presents greater clinical and infection control challenges.⁹ Furthermore, the recent outbreak of vancomycin-resistant or glycopeptide-resistant S aureus infections in a French ICU was difficult to control, and costly, and may be a harbinger of future problems, especially with the recent reduction in vancomycin sensitivity from a minimum inhibition concentration of 4 to 2 μg/mL.¹⁰ Control of MRSA will require more aggressive antibiotic control, focused on reduced use of antibiotics, such as fluoroquinolones and improved infection control. Infection control for high-risk populations and certain health-care facilities may require improved screening, isolation, and eradication of MRSA,¹¹¹²¹³ such as the “search and destroy” strategy recently outlined by Vos and coworkers.¹⁴

Bacterial colonization of the aerodigestive tract and entry of contaminated secretions into the lower respiratory tract are critical in the pathogenesis of VAP and major targets for prevention (Fig 1 ).¹²¹⁵¹⁶ The endotracheal tube is a major risk factor for VAP, as it permits leakage of oropharyngeal secretions around the cuff and may act as a nidus for the growth of intraluminal biofilm.

Prevention efforts targeting VAP must be part of an evidence-based, multidisciplinary prevention program that has a “core” team with an agenda focused on patient safety and quality improvement (Fig 2, 3 ).¹⁶¹⁷¹⁸ Optimally, the team should be led by a “champion” of the cause and include interested clinicians, respiratory care staff, administrators, risk management staff, and other stakeholders as core team members (Fig 3). The responsibilities of this group include setting prevention benchmarks, establishing goals and time lines, and providing appropriate education and training, audits, and feedback to the staff, while continually updating themselves on the relevant clinical and prevention strategies.

Prevention programs should be “marketed” to hospital administrators and others involved in resource allocation by demonstrating that preventing VAP results in improved clinical outcomes and significantly reduced costs. Data from Rello in 2002, demonstrated that on average, a case of VAP increased hospitalization by 12 days, mechanical ventilation by 10 days, ICU stay by 6 days, and hospital costs by $40,000; similar results have been reported in a suburban hospital by Warren and coworkers.³¹⁹ In addition to these direct savings, the growing trend of public reporting of institution-specific infection rates and other outcome data are not only increasing but may eventually effect hospital reimbursement rates.

As with other prevention efforts, interventions aimed at reducing VAP should focus on evidence-based interventions, for which efficacy and cost-effectiveness have been clearly supported by clinical studies and experts in the field (Table 1 ).¹²¹⁷ Initially, it may be more prudent to focus on a limited number of feasible, cost-effective prevention strategies for VAP prevention. In the Institute for Healthcare Improvement (IHI) 100,000 Lives Campaign, hospitals are challenged to adopt as many of the six recommended initiatives to reduce health-care–associated infections.¹⁸ The VAP or “ventilator bundle” initiative includes five simple components: elevation of the head of the bed to between 30° and 45°, a daily “sedation vacation,” daily assessment for readiness to extubate, and prophylaxis for peptic ulcer disease and deep vein thrombosis. Some participating hospitals using this approach are reporting zero episodes of VAP over sustained periods of time (Donald Berwick, MD; IHI National Forum; personal correspondence; December 13, 2005). Confirmation of these dramatic results in peer-reviewed journals is eagerly anticipated.

Staff education, particularly targeting those clinicians and staff who manage patients receiving mechanical ventilation, is a cornerstone for efforts to reduce the incidence of VAP. Kollef¹⁷ and Babcock and coworkers¹⁹ initially reported the success of a VAP educational prevention program carried out in five ICUs. The program, developed by a multidisciplinary team, targeted respiratory care providers and intensive care nurses who completed a self-study module on risk factors for VAP at baseline and after the program interventions. Relevant in-service teaching programs were coordinated with staff meetings, and fact sheets and posters were placed in the ICU and respiratory care departments. Rates of VAP dropped nearly 58%, to 5.7/1,000 ventilator days, and cost savings were estimated to be from $425,606 to $4,000,000. Babcock et al,²⁰ using an extension of this program in an integrated health-care system involving four hospitals, reported a 46% reduction in VAP over an 18-month period.

Perhaps one of the most important and underappreciated prevention strategy is adequate staffing, particularly in critical care units.¹¹⁶ Staffing must be sufficient to allow patient care to be provided while ensuring that staff are able to comply with essential infection control practices and other prevention strategies.¹⁹²¹

In a study of abdominal aortic surgery patients by Dang et al,²² decreased nursing staffing was associated with significantly higher rates of respiratory and cardiac complications than in patients who had higher intensity nursing. Currently, this is of critical importance due to severe nursing shortages and staffing reductions due to budget constraints. Nurse-to-patient ratios should be 1:1 for high-risk complicated ICU patients, or 2:1 for patients with lower disease acuity. Currently, efforts to establish legislation that would cap the number of patients per nurse are underway in some states.

Infection control programs have repeatedly demonstrated efficacy in reducing infection rates and in controlling the spread of MDR organisms.^12151721 Unfortunately, staff compliance with proven infection control measures, such as hand disinfection, is often poor and inconsistent. Staff education aimed at infection control must be inclusive, frequent, and reiterative. Special attention must be directed to house staff, students, volunteers, and visitors who may not be included in regularly scheduled infection control educational programs.

Surveillance of ICU infections to identify and quantify endemic and new MDR organisms with data feedback is a critical. Communication of current data among clinicians, laboratory, pharmacy, and infection control staff is essential.

Organism-specific strategies for specific MDR pathogens are recommended.^91011121314 For MRSA, vancomycin-resistant S aureus, or glycopeptide-resistant S aureus isolates, more aggressive screening, and isolation are recommended, and more aggressive eradication has been advocated.^10111314

Antibiotic control programs are also extremely important in the overall effort to control infections, reduce emergence of MDR organisms, and control spiraling health-care costs. For example, reduced use of fluoroquinolones has been associated with reduced rates of MRSA infection.¹² Antibiotic control strategies are complicated, and should be focused, dynamic, carefully monitored, and may vary by type of MDR pathogen. For example, control of specific types of MDR Gram-negative bacilli, may require “squeezing the balloon at multiple sites” to prevent the emergence of other MDR pathogens, as nicely summarized Rahal and coworkers.²³ In addition, an infectious disease pharmacist for the ICU team or computerized surveillance programs to target interventions and aid in determining optimal drug regimens should be considered.¹² Data from antibiotic rotation programs are more difficult to evaluate, but this approach has been advocated for reducing MDR pathogens.²²⁴

Most cases of VAP, particularly those caused by resistant bacteria, are the result of pathogenic microorganisms in the host and environment.¹²²¹ Although it is widely appreciated that the hospital environment is swarming with microorganisms, this does not necessarily translate into nosocomial infections and widespread routine environmental sampling is not recommended. Studies²¹²⁵ are beginning to implicate the inanimate environment as an indirect contributor to nosocomial acquisition of some potential pathogens. Special interventions, including targeted environmental sampling and more aggressive environmental disinfection, may be indicated during nosocomial outbreaks, particularly those involving MDR organisms or organisms that are more resistant to routine cleaning.²⁵ Legionella species can be recovered from 12 to 70% of hospital water systems and this source of nosocomial outbreaks remains underappreciated.¹²¹

In contrast to rotational beds, semirecumbent patient position is a low-cost, easily accessible intervention, and may be a more practical and more tolerable approach than rotational beds or prone body position.²⁶ Maintaining patients who are receiving mechanical ventilation or who are enterally fed in a 30° to 45° semirecumbent position, particularly during enteral feeding, continues to be strongly recommended based on the VAP reduction in one randomized study.¹²¹⁷

A more recent study by van Nieuwenhoven et al,²⁷ in which patients receiving mechanical ventilation were randomly assigned to backrest elevation of 45° vs the standard of 10°, demonstrated barriers to implementing this strategy. Backrest elevation was measured continuously during the first week of ventilation with a monitoring device. The targeted backrest elevation of 45° was not reached; the actual achieved difference was 28° vs 10°, which did not reduce VAP. Similarly, Grap and Munro²⁸ monitored patient position in ICU patients using a bed frame elevation gauge or electronic bed readout and found very low compliance with maintaining semirecumbent patient position, with a mean backrest elevation of only 19.2° with 70% of subjects maintained in a supine position. Maintaining patients receiving mechanical ventilation or who are enterally fed in a semirecumbent position may need to evaluate more realistic targets.

Hypergylcemia, relative insulin deficiency, or both may directly or indirectly increase the risk of complications and poor outcomes in critically ill patients.²⁹ Initial randomized studies of intensive insulin therapy to maintain blood glucose levels between 80 and 110 mg/dL in surgical ICU patients had lower overall mortality (4.6% vs 8%, p < 0.04) and was greater for those with ICU stays > 5 days (10.6% vs 20.2%, p = 0.005), coupled with a 46% reduction of bloodstream infections, decreased acute renal failure, fewer antibiotic treatment days, and significantly shorter length of mechanical ventilation and ICU stay.

To address risk-benefit issues, Egi and colleagues,³⁰ using a retrospective outcome study of intensive insulin therapy (ITT), reported that hypoglycemia in SICU patients varied from 1.4 to 2.7%, and estimated that the number of patients needed to be treated to save one life varied from 38 to 113, whereas the rate of hypoglycemia (number needed to harm) varied from 7 to 13 patients.

There were also questions on the generalizability of ITT to medical ICU patients. In a recent randomized study³¹ of 1,200 medical ICU patients, ITT did not significantly reduce hospital mortality overall, and increased mortality in patients with ICU stays < 3 days. However, the ITT group had reduced acquired renal failure, duration of mechanical ventilation, and length of ICU and hospital stay. Difficulty in predicting length of stay is difficult; concerns about the risks of hypoglycemia, resource implications, and assessing the benefit of ITT in different hospitals require further evaluation.

Enteral feeding is preferred to parenteral feeding, but aspiration pneumonia is a complication.^16303233 Bowman and coworkers³³ instituted an evidence-based, enteral feeding protocol in which 78 to 85% of patients reached their enteral feeding goal and aspiration pneumonia rates decreased from 6.8 to 3.2/1,000 patient days. Such protocols should be reviewed by multidisciplinary committees to standardize enteral nutrition protocols and risk reduction for VAP.

Early gastrostomy for enteral feedings has been suggested as a strategy to reduce VAP in patients with head injury and stroke.³⁴ In a small, randomized clinical trial³⁴ of 20 patients with gastrostomy vs 21 control subjects, rates of VAP were reduced (10% vs 38%, respectively), and most of the VAP cases were late-onset VAP (> 5 days). Further studies with higher enrollment are needed to further assess this strategy in these high-risk patients.

Histamine type-2 (H₂)-agonists and antacids have been identified as independent risk factors for ICU-acquired HAP. Sucralfate has been used for stress bleeding prophylaxis, as it does not increase intragastric acidity or gastric volume but is less effective in preventing GI bleeding.¹²⁴ Bornstain and workers³⁵ examined risk factors for early onset VAP (from 3 to 7 days) in 747 patients. Several different variables were identified in the univariate analysis, but only sucralfate used in the first 48 h of ICU stay and unplanned extubation were predictors of VAP in the multivariate analysis; antibiotics were protective. Although the dose of sucralfate was not mentioned, and the mechanism is unclear, an increased risk of VAP was also reported in an earlier study³⁵ of patients with ARDS. Data indicate that H₂-blockers or proton-pump inhibitors (PPIs) are associated with lower rates of GI bleeding when compared to sucralfate, which may be important as transfusion is also a risk factor for VAP.¹²

Although transfusion was suggested as a risk factor of nosocomial infection and a modifiable risk factor for VAP in the American Thoracic Society/Infectious Diseases Society of America guideline, in a secondary analysis from a recent, large study³⁶ of transfusions, it was identified as an independent risk factor for VAP. This may become a more important modifiable risk factor, as recent data from Levy and coworkers³⁷ reported that patients receiving mechanical ventilation received transfusions at a higher pretransfusion hemoglobin level than patients not receiving mechanical ventilation (8.7 vs 8.2, p < 0.0001).

Oral care has been recommended to prevent VAP in several studies. In a recent study, Mori and coworkers³⁸ compared rates of VAP in a nonrandomized group compared with historical control subjects. Incidence of VAP in the oral care group was 3.9 episodes/1,000 days vs 10.4 episodes in the control group. Although there are concerns about the study design, oral care has intuitive benefits and limited cost.

The GI tract, VAP, and the clinical evidence for the efficacy of selective decontamination of the digestive tract (SDD) were recently review by Kallet and Quinn.³² Furthermore, Liberati and coworkers³⁹ published an extensive review of antibiotic prophylaxis to reduce respiratory tract infections in adults receiving intensive care. The authors³⁹ conclude that for topical and systemic antibiotic prophylaxis, 5 patients would need to be treated to prevent one infection and 21 patients would need to be treated to prevent one death. No recommendation was made for topical prophylaxis. In a recent large study of SDD by de Jonge and coworkers⁴⁰ in 2003, SDD was highly effective with no increased antibiotic resistance observed. Citing concerns over rapid increases in antimicrobial resistance in the hospital setting, coupled with the association between MDR pathogens and poorer patients outcomes and the dearth of new antimicrobial agents in the pipeline, recent guidelines¹²⁴¹ have suggested that SDD should considered for selected ICU populations and clinical scenarios but not be employed “routinely” for VAP prevention.

Topical antiseptics, such as chlorhexidine, provide an attractive alternative, but the initial reported success in cardiac surgery patients could not be confirmed by other studies. Koeman and coworkers⁴² provide further important data from a multicenter, double-blind, randomized, clinical trial of VAP outcomes for subjects treated with 2% chlorhexidine paste vs patients randomized to 2% chlorhexidine plus 2% colistin paste to provide greater activity against Gram-negative bacilli compared to placebo. Compared to the placebo group, the daily risk of VAP was reduced by 65% in the chlorhexidine group (p = .01) and 55% in the chlorhexidine-colistin group (p < 0.03). This impressive result for an inexpensive, nontoxic, topically applied modality warrants further attention but is difficult to reconcile with the absence of effect on ventilator-days, length of stay, or mortality. It is important to measure how prophylactic use of chlorhexidine and chlorhexidine-colistin complement other effective prevention strategies, and resistance could become an important issue over time.

Iseganan, a topical antimicrobial peptide, active against aerobic and anaerobic Gram-positive and Gram-negative bacteria and yeasts, was evaluated in a randomized, double-bind trial⁴³ to prevent VAP. Although there was a significant reduction in colonization in the treatment group, the rate of VAP among survivors (16% vs 20%) and 14-day morality was similar (22% vs 18%). Although protegrins are ubiquitous antimicrobial peptides, and in human trials were able to reduce oral colonization by two logs, these results raise several questions about iseganan efficacy and why it failed.

Interesting data from a double-blind pilot study of subjects randomized to oral and gastric lactobacillus GG (2 × 10⁹ colony forming units bid) [n = 19] vs placebo (n = 21) were presented by Morrow and coworkers at the 2005 American Thoracic Society Annual Meeting Symposium in San Diego. Demographics and acute physiology and chronic health evaluation (APACHE) II scores were similar in the two groups, but the lactobacillus GG group had significantly higher rates of colonization with normal flora (p < 0.03), fewer oral pathogens, less clinical VAP (26% vs 45%. p = 0.2), lower microscopic VAP (11% vs 33%, p = 0.08), and a lower mortality rate (0% vs 10%, p = 0.17). Further studies are in progress.

Noninvasive positive pressure ventilation (NPPV) provides ventilatory support without the need for intubation and for earlier removal of the endotracheal tube to reduce complications related to prolonged intubation. Burns and coworkers,⁴⁴ in a recent Cochrane review, reported significant benefits: decreased mortality (risk ratio [RR], 0.41; 95% confidence interval [CI], 0.22 to 0.76), lower rates of VAP (RR 0.28; 95% CI, 0.0.90 to 0.85); decreased length of ICU and shorter hospital stays; and lower duration of mechanical support. The impact of NPPV is greater in patients with COPD exacerbations or congestive heart failure than for patients with VAP. Recent data also indicate that NPPV may not be a good strategy to avoid reintubation after initial extubation, and is recommended for hospitals with staff who are experienced in this technique.⁴⁵

The endotracheal tube lumen is also a nidus for the growth of bacteria-encased in biofilm. Rates of bacterial biofilm formation increase over time, are protected from host humoral and cellular defenses, and antibiotics, and may contain high concentrations of bacteria.¹⁴¹⁶⁴⁶ Biofilm-encased bacteria also are less susceptible to killing by host defenses. Suctioning of patients or passage of a bronchoscope may dislodge biofilm-encased bacteria that may increase the risk of late-onset VAP.

Prevention of bacterial biofilm formation on urinary catheters has been reduced by the use of a silver coating. The use of a silver-coated endotracheal tubes, which was effective in preventing VAP in a dog model, is currently being evaluated in a large, multicenter, randomized clinical trial of patients receiving mechanical ventilation.

Moller and coworkers,⁴⁷ using a retrospective study design, examined the potential benefit of early tracheostomy ET (< 7 days) vs late tracheostomy in severely injured surgical SICU patients. Patients with late tracheostomy had significantly higher rates of VAP (42.3% vs 27.2%, p < 0.05), duration of mechanical ventilation, and length of ICU stay. The authors⁴⁵ suggest that if patients will require prolonged ventilation (> 7 days), that tracheostomy be performed between day 3 and 7. In a trial of 60 trauma patients randomized to early tracheostomy by Barquist and coworkers,⁴⁸ the study was terminated as the intervention had no effect on mortality, rates of VAP, ICU stay, or other outcomes. A systematic review and metaanalysis by Griffiths and coworkers⁴⁹ from 406 patients in five studies also reported no reduction in pneumonia, mortality, ventilator days, or length of ICU stay.

Efforts to reduce acute lung injury by using smaller tidal volumes and lower pressures have been suggested along with sedation vacations to facilitate weaning.¹²¹⁷⁵⁰ Recently, Schweickert and coworkers⁵¹ evaluated seven complications in 128 patients receiving mechanical ventilation and continuous infusions of sedative drug who were randomized to daily interruption of sedative infusions (n = 66) vs sedation directed by the medical ICU team without this strategy (n = 60). Daily interrupted sedative infusions reduced ICU length of stay (6.2 days vs 9.9, p < 0.01), duration of mechanical ventilation (4.8 vs 7.3 days, p < 0.003), and the incidence of complications per patient (13 complications in 12 patients vs 26 complications in 19 patients, p < 0.04).

Weaning protocols are recommended to limit the duration of mechanical ventilation. Dries and coworkers,⁵² using a standardized weaning protocol, reduced days of mechanical ventilation (ventilator days/ICU days) from 0.47 to 0.33, numbers of patients failing ventilation (25 vs 43), and reduced rates of VAP (15% to 5%). Although there are a number of confounding variables with the study design, efforts to remove the endotracheal tube without reintubation should be encouraged.

Continuous aspiration of subglottic secretions (CASS), through use of specially designed endotracheal tubes with wider, elliptic holes, helps facilitate drainage.⁵³ In a recent metaanalysis,⁵⁴ CASS reduced the incidence of VAP by half (RR, 0.51; 95% CI, 1.7 to 2.3), shortened ICU stay by 3 days (95% CI, 2.1 to 3.9), and delayed the onset of VAP by 6 days. CASS was also cost-effective, saving $4,992/case of VAP prevented or $1,872/patient, but mortality was not affected. However, when CASS was combined with semirecumbent positioning, no clinical benefit was observed, which underscores the importance of interactive prevention strategies.⁵⁵

Ventilator circuit issues and methods of humidification in relation to VAP were recently summarized by Branson.⁵⁶ Frequency of circuit changes does not prevent VAP and is an area for substantial cost saving. Condensate collecting in the ventilator circuit can become contaminated from patient secretions or by opening the circuit; vigilance is needed to prevent inadvertently flushing the condensate into the lower airway or in-line nebulizers at the bedside or during patient transport. Metered-dose inhalers may be safer for the delivery of bronchodilators than nebulizers, which if contaminated, may produce bacterial aerosols.

There have been conflicting reports on the use and benefits of heat/moisture exchangers (HMEs) compared to heated humidifiers for preventing VAP.⁵⁶ A recent metaanalysis by Kola and coworkers⁵⁷ demonstrated a reduction in the relative risk of developing VAP in the HME group (relative risk, 0.7; 95% CI, 0.5 to 0.04) but may have been affected by the large difference in the outcomes in one of the studies. For patients with a mean ventilation duration > 7 days, the relative risk for VAP fell to 0.57 in the HME group (95% CI, 0.38 to 0.83). A more recent, large, randomized study by Lacherade and coworkers⁵⁸ found no benefit for the HME group. In another study⁵⁴ of HMEs using historical control subjects, patients who received mechanical ventilation > 2 days reported a significant reduction in VAP (p = 0.01).

The focus on prevention is focused on ICU patients, but these patients are at increased risk for relapse or reinfection during their rehabilitation. Therefore, efforts should be directed at risk reduction at discharge, such as routine vaccinations and patient education (Table 1).

Despite rapid technological and treatment advances in medicine, dramatic reductions in rates of VAP and effective use of complex prevention and management guidelines remain elusive.¹² Prevention outcomes are directly related to reducing risk (Fig 4 , Table 1). Prevention involves planting a tree, nurturing it, pruning it, and watching it grow and spread seeds for more trees. Investing in prevention can pay great dividends in terms of improved quality of life, morbidity, and mortality. In addition, prevention can have a huge impact in reducing length of stay and health-care costs during acute care. Spreading the seeds of prevention into long-term care and rehabilitation facilities is also vitally needed.

As described in a recent commentary by Berwick et al,¹⁸ the laudable goal set forth by IHI to reduce deaths among hospitalized patients in the United States by 100,000 over 18 months by improving patient quality and safety set a very high bar.¹⁷ Each of the six “100,000 Lives Campaign Interventions,” which includes VAP, is conceptually simple and feasible; notably, each strategy in the “VAP bundle” is not new or expensive.

The IHI 100,000 Lives Campaign may be the call to action that is needed to disseminate prevention and safety information, and implement prevention guidelines consistently and broadly.¹⁸ Even if the campaign falls short of its goal of saving 100,000 lives, the enrollment of over half of the hospitals in the nation in this effort can provide a valuable infrastructure for sowing seeds, planting trees, and measuring outcomes. This infrastructure, coupled with endorsements of government agencies (Joint Commission on Accreditation of Healthcare Organizations and Medicare), and medical, nursing, and public health groups, translates into powerful lobbies for the advancement of patient safety, quality care, and the necessary resources to incorporate prevention into practice.⁵⁹

I thank Kathleen A. Steger Craven for her assistance in preparing, writing, and editing this manuscript. Her experience, insights, ideas, and contributions are greatly appreciated.