Selective Digestive Tract Decontamination (SDD)
Question 14 from the second paper of 2013 and the identical Question 22 from the second paper of 2017 asks for a substantial amount of detail regarding SDD. A quick answer to "critically evaluate the role of SDD" can be found in the Discussion section there.
Below is a summary of recent evidence for SDD, scraped together from the following sources:
Oh's Manual mentions SDD only in brief, in the context of preventing liver transplant complications, and preventing bacterial translocation from the gut in severe acute pancreatitis.
Rationale for decontamination of the gastrointestinal tract
The whole point of SDD is to prevent nosocomial infection caused by the translocation of gastrointestinal microbes out of the GI tract and into areas where they dont belong.
- Critical illness causes an overgrowth of normal and opportunistic flora.
- Specifically, the organisms which enjoy a population explosion are exactly those which are known to cause common nosocomial infections: predominantly oropharyngeal and upper GI organisms.
- Increased population of this flora leads to the proliferation of multiple bacterial clones, which is an ideal environment for developing antimicrobial resistance.
- Thus, it stands to reason that if one were to eradicate these organisms, one would reduce the incidence of nosocomial infection in critical illness, and reduce the incidence of antimicrobial resistance.
- There is no agreed-upon protocol.
- Usual antibiotics are the "PTA" cocktail:
- polymyxin E
- amphotericin B
- The SuDDICU authors have complained that these days its harder and harder to get "good old fashioned amphotericin" and the more recent trials have used nystatin instead, which is a poorly absorbed polyene antifungal (i.e. from the same class), cheap and already enjoying universal popularity as a treatment for oral candidiasis.
- This goop is applied to the oral cavity, and injected down the NG tube.
- Vancomycin can be added if there are high local rates of colonisation
- A controversial step is the addition of a short course of IV antibiotics.
- This is usually a broad spectrum cephalosporin eg. cephotaxime, and it is used for 4 days only. Many study protocols omit this.
- Surveillance cultures of the throat and rectum are recommended
The proposed benefits of SDD:
The beneficial effects are expected to manifest in the following ways:
- Decreased incidence of VAP.
- Decreased incidence of pseudocyst infection in acute pancreatitis.
- Decreased systemic infection rate in cases of gastrointestinal haemorrhage.
- Decreased incidence of spontaneous bacterial peritonitis in patients with ascites.
- Protection against early gram-negative sepsis following liver transplantation
Expected disadvantages of SDD
- May result in increased rates of antimicrobial resistance
- Expensive in terms of antibiotic doses, staff workload, and processing of surveillance samples
- Local flora may not respond to standard antibiotic cocktails
Evidence for and against SDD
- To date, there are numerous RCTs - LITFL counted 60, with over 15,000 patients. A recent (2013) meta-analysis has identified 64 studies.
- The studies have revealed that
- There is a mortality benefit (OR = 0.73, NNT = 18)
- Specificaly, the largest trial to date (de Smet, 2009) found an all-cause mortality reduction by 3.5%, from an absolute rate of 27.5% down to 25%.
- Length of stay is decreased
- VAP incidence is decreased by 72%
- The "full-scale" SDD protocol (with a short course of IV antibiotics) is superior to the "limited" SDD where only the oropharynx is decontaminated (and both strategies are superior to oral chlorhexidine)
- A 2004 Cochrane review found that whereas the "full-scale" SDD decreases the incidence of respiratory tract infections and mortality, the use of oropharyngeal decontamination alone had no effect on mortality- only on incidence of pneumonia.
Critique of supportive evidence
- The studies all suffer from heterogeneity and poor experimental design.
- Outcome definitions vary considerably, especially as there is no universally accepted definition of VAP.
- Most of the studies were conducted in countries with a lower local MRSA prevalence (Germany and the Netherlands). In these affluent Norther European nations, rates of multiresistant organism colonization would not be representative.
- Placebo solutions may have been microbial vehicles in these studies, increasing the incidence of VAP among the control group.
- There are no good studies of the effect of SDD on the development of multi-resistant organisms (but poor-quality studies suggest that there is no effect).
Support for SDD
- There is little support for this practice among the Australian intensivist community.
- The practice has been regarded with scepticism in Australian ICUs, and there has been no broad department-level support for it here. Opponents cite paucity of good quality evidence and non-generalisability of European data in local intensive care units.
- The Surviving Sepsis guidelines have recently favoured this practice with a 2B-level recommendation; however, there is also little support for the Surviving Sepsis guidelines among the Australian intensivist community.
- In contrast, the Dutch community has embraced the SDD protocols after the de Smet study (NEJM, 2009) and apparently the practice of selective oral decontamination is now the standard of care in Dutch ICUs.
With such criticism levelled at the current (extensive) body of evidence, one wonders what sort of multiplanetary interspecies triple-blinded trial would be required to defeat the doubts of the sceptics.
One such trial is currently cooking - it is the SuDDICU trial by the CCCTG.
Of it, one of the principal investigators, Ian Seppelt, writes for the Crit IQ blog:
"... Our goal is to do not just the next (37th) RCT for the next Cochrane metaanalysis, but to do the DEFINITIVE trial which answers the question, one way or the other...we will be using a mixture of nystatin, colistin and tobramycin orally and enterally... we will need up to 50,000 patients over a number of years."
This has great promise, because it could potentially represent a substantial mortality benefit from the use of a fairly simple (and cheap) routine measure. Time will tell.