Disaggregating the Benefits of Ventilator Bundle Components

There is a very nice study by Michael Klompas and colleagues in September's JAMA Internal Medicine. The team sought to disentangle the benefits or harms of the individual components of current ventilator bundles including: head-of-bed elevation, sedative infusion interruptions, spontaneous breathing trials, thromboprophylaxis, stress ulcer prophylaxis, and oral care with chlorhexidine.

Prior studies had found potential harm associated with stress ulcer prophylaxis (pneumonia) and oral chlorhexidine (higher mortality). This same group published a meta-analysis that called into question the benefits of routine oral care with chlorhexidine (CHG) in ventilated patients. For this retrospective cohort, they examined the associations between exposure to individual ventilator bundle components on a day-by-day basis and ventilator-associated events (VAEs), duration of mechanical ventilation, ventilator mortality, hospital length of stay, and hospital mortality.

The cohort included 5539 consecutive patients who were exposed to mechanical ventilation for at least 3 days. They measured the association between individual process measures and VAEs using Cox proportional hazards regression models with fixed and time-varying covariates and censored patients on extubation or death, whichever came first. Interestingly, they calculated hazard ratios for each bundle component "as the contrast between 4 days of continually performing the process measure vs 4 days of not doing so", since they wanted to allow for the possibility that process measures might have an immediate or delayed effect on each outcome.

I have included the patient-outcomes table below. The most interesting finding, among many tested associations, was that oral care with chlorhexidine was associated with an increased risk for ventilator mortality (HR, 1.63; 95% CI, 1.15-2.31; P = .006). In another table, they reported that stress ulcer prophylaxis was associated with an increased risk for possible ventilator-associated pneumonia (HR, 7.69; 95% CI, 1.44-41.10; P = .02).

This was a very thoughtfully completed and written study - I encourage you to read it (and the accompanying invited editorial) beyond my quick overview. The authors concluded: "we should revisit the classic ventilator bundle. Possible revisions include... a reappraisal of whether oral care protocols should be revised to exclude chlorhexidine therapy, and the reservation of stress ulcer prophylaxis for patients at marked and immediate risk for upper gastrointestinal tract bleeding rather than prescribing them for all patients undergoing ventilation."  It will be interesting to see how slowly these recommended changes occur...

Bundle fumble?

This week's JAMA has an excellent review (free full text here) on the prevention of ventilator-associated pneumonia (VAP). Specifically, the authors offer a critical assessment of the widely utilized IHI VAP bundle. They offer two important conclusions:

• "The ability of the bundle to prevent VAP has not been definitively established with high quality studies."
• "No large randomized study has demonstrated that reducing VAP using any strategy, including the IHI bundle, is associated with improvements in clinical outcomes."

Here's another example where an intervention has been touted as gospel, pushed hard and implemented broadly without the evidence necessary for a hardline approach.

Photo: OregonLive

Will this contraption work?

One of the biggest obstacles to the accurate diagnosis of pneumonia is determining whether an expectorated sample is representative of the upper airway/oral cavity (representing colonization or, at most, tracheobronchitis) or the lower respiratory tract (representing a true pathogen).

A group from the Georgia Institute of Technology recently developed a gadget they claim can separate lower respiratory from upper respiratory aerosols. The report can be found here, in the Journal of Medical Devices.

Tracheotomy, VAP, p-values and death

There is a new RCT just published in JAMA by a large group in Italy looking at the benefits of early (day 6-8) vs late (day 13-15) tracheotomy completed in 12 ICUs. The primary endpoint was VAP. There is also a very nice accompanying editorial. There are several interesting findings. First, patients randomized to early tracheotomy were less likely to develop VAP by day 28, 14% vs 21%, but the p-value was 0.07. Since the p value was greater than 0.05, the authors were forced to say that there was no benefit from early tracheotomy.

Interestingly they also found significantly greater vent-free days, ICU-free days, successful weaning and ICU discharges in the early tracheotomy group. There was even a trend towards higher survival in the early vs late group, HR=0.80, 95% CI 0.56-1.15. The authors and editorial do a nice job of pointing out that 31% of early and 43% of the late group didn't even receive a tracheotomy due to impending extubation or death. The editorial even makes the point that selecting an early tracheotomy is really a strategy of more trachs. The study did not assess patient comfort, which may be associated with early tracheotomy.

What is always troubling to me is that scientists, editorialists, journals and clinicians are stuck in this p-value trap. Here we have a study, a very good randomized trial, which shows likely clinically significant reductions in VAP and potentially lower mortality, but since the study was underpowered we are forced to say "no difference." I wonder if you calculated how many patients are intubated each year in the US (or Italy) and reduced VAP rates by 33%, how many VAPs would be prevented and how many deaths would be prevented? I know this study should be repeated, but will it? You have a negative JAMA study, what's the incentive? I describe this phenomenon as "Death by p-value."

VAP: Do you know it when you see it? (Again!)

There's a new paper on the utility (or lack thereof) of CDC's definition of ventilator-associated pneumonia. In this study 4 persons reviewed 50 cases of ventilated patients with respiratory deterioration >48 hours after intubation. Two reviewers were experienced infection preventionists who applied the CDC definition. A third IP used a modification of the CDC definition that was more quantitative, and the fourth reviewer was a physician board-certified in infectious diseases and critical care who used clinical judgment to define VAP. Using the standard definition, one IP assigned the VAP diagnosis to 11 patients and the other, 20 patients. The IP using the modified definition assigned 15 cases as VAP. The physician diagnosed VAP in 7 patients. The IPs agreed on 62% of cases (kappa=0.40). All 4 reviewers agreed on the VAP diagnosis in only 4 cases. This is not the first study to show how complex assigning the diagnosis of VAP can be.

Given that public reporting has raised the stakes to high levels, the CDC can no longer ignore this issue. Consumers cannot make choices on where to receive care if inter-hospital comparisons of infection rates are not valid. There needs to be a convening of IPs and hospital epidemiologists who use the definitions on a daily basis to thoroughly assess each of the HAI case definitions and begin to work on the development of new ones that will be fair to hospitals and helpful to consumers. Otherwise, it's garbage in, garbage out, and the entire concept of public reporting is undermined.

Defining our way to zero?

Here is a quote from an Emerging Infections Network post today, regarding how we define central-line associated bloodstream infections (CLABSI):

Is it really worthwhile to adjudicate blood stream infections in patients with central lines as "primary" or "secondary"? The adjudication according to the "definition" is still subjective. For example, a Klebsiella bacteremia in a pt with a PICC and PEG was ascribed to "gastroenteritis" since the pt had some coincident diarrhea and stool fecal leukocyte+ (hence bacteremia was considered secondary). A candidemia was attributed to pneumonia since the pt was immunocompromised (on steroids for BOOP), had a fluctuating CXR and had Candida in his sputum (hence the candidemia was considered secondary). I was told that these attributions were completely reasonable since they were compatible with the definitions and that the institution regularly passes muster when audited. It is distasteful to argue but anyone looking closely would see a discrepancy between the clinical diagnosis and the adjudicated diagnosis. Since the public is taking these numbers seriously, there is a problem.

This EIN post goes to the heart of a very important issue in healthcare associated infection reporting—the subjective interpretation of National Healthcare Safety Network (NHSN) definitions. Of course we already know that ventilator associated pneumonia (VAP) rates are a load of crap (I’ll leave it to Klompas and Platt to explain why). It is less well recognized how much fiddling is going on with the CLABSI definition. The post above is a great example of what is happening across the country in hospitals that are under increasing pressure to “get to zero”, and as public reporting of infection rates becomes the norm rather than the exception. Hence those hospitals that apply the NHSN CLABSI definition very strictly are punished with higher CLABSI rates. Meanwhile, hospitals celebrating “zero” rates may in fact be no more “safe” than before they began fudging their definitions.

How to fix this? Either via an expensive and cumbersome validation system for public reporting (any ideas for CDC and state public health departments on how to do this?), or via more specific definitions from NHSN.

Decontamination: not so selective?

Remember the NEJM study from the Netherlands that compared selective digestive decontamination (SDD) vs. selective oropharyngeal decontamination (SOD) vs. placebo for infection prevention in ICU patients? The upshot, using 28 day mortality as an endpoint, was that both SDD and SOD were beneficial, but there was no evidence that one was better than another.

These investigators have now published a report on changes in antimicrobial resistance after the prophylactic use of antibiotics in this study (tobramycin, polymyxin E and amphotericin for oropharyngeal (SOD) or nasogastric (SDD) administration, and 4 days of IV cefotaxime (SDD)).

Surveillance of rectal and respiratory tract samples from patients in the 13 participating ICUs demonstrated that resistance to ceftazidime, tobramycin, and ciprofloxacin increased in GI tract flora after the SDD intervention and increased in respiratory flora after both SDD and SOD interventions.

I have to read the fine print more closely, but this confirms my view that we should stick with chlorhexidine oral care to suppress oropharyngeal flora, rather than SDD or SOD approaches that use therapeutic antibiotics.

VAP: Do you know it when you see it?

Ask any infection control nurse, infectious diseases physician, or intensivist and they will tell you of the difficulty in trying to define and diagnose ventilator-associated pneumonia (VAP). This is because there are a number of other conditions that can mimic VAP. A new study in the Journal of Critical Care compares three VAP definitions to autopsy findings in 253 patients to determine the utility of the definitions. The CDC definition was not used, but is similar to one of the definitions applied in the study. None of the definitions performed well. The most sensitive definition (65%) had a specificity of 36%. On the other hand, one definition had a nearly perfect specificity (99%), but its sensitivity was dismal at 5%. The results of this study coupled with conventional wisdom should be taken as a warning that public reporting of VAP rates may mislead consumers, and other metrics for reporting should take precedence.