PedsCCM: The Pediatric Critical Care Website:
The multidisciplinary and practical resource for Pediatric Critical Care on the Internet Peer reviewed original content integrated with links to selected web resources ... now 19 years old... thanks to everyone for your great support!..


Criteria abstracted from The Users' Guides to the Medical Literature series in JAMA

Survival benefit in critically ill burned patients receiving selective decontamination of the digestive tract: a randomized, placebo-controlled, double-blind trial.

de La Cal MA, Cerda E, Garcia-Hierro P, van Saene HK, Gomez-Santos D, Negro E, Lorente JA.

Ann Surg. 2005 241(3):424-30.[abstract]

Reviewed By: H Rowlands MD, Birmingham Childrens Hospital, Birmingham UK

Review posted September 22, 2006

I. What is being studied?:

1. The study objective:

  • Does Selective Decontamination of the Digestive Tract (SDD) reduce mortality from any cause in patients with severe burns?
  • Does SDD reduce the incidence of pneumonia in patients with severe burns?

2. The study design:

  • Prospective.
  • Randomised.
  • Double blind.
  • Placebo controlled.
  • Sample size calculated for a decrease in the incidence of pneumonia by 60%, 80% power. n = 102.
  • 6 bed burns Intensive Care Unit (ICU) in a tertiary hospital (80 admissions per year).

3. The patients included:

  • 14 years old or older.
  • 20% or greater burns total body surface area (total BSA) and or inhalational injury diagnosed bronchoscopically.
  • 3 or less days interval between injury and ICU admission

4. The patients excluded:

  • Duration of stay in ICU less than 3 days.
  • Withdrawal of treatment at less than 3 days.
  • Immunosuppression.
  • Pregnancy.
  • Inhalational injury not needing ventilation in first 3 days.

5. The interventions compared:

Study Group received:

1. Intravenous cefotaxime 1g three times a day for 4 days.

2. Four times daily application of 0.5g topical oropharyngeal paste containing non-absorbable polymyxin E, tobramycin and amphotericin B.

3. Four times daily digestive administration of 10ml of solution containing 100mg polymyxin E, 100mg tobramycin and 500mg amphotericin B.

4. Surveillance samples throat and rectum on admission and twice weekly.

Controls received:

1. iv 0.9% saline

2. placebo paste

3. placebo digestive solution.

4. Surveillance samples throat and rectum on admission and twice weekly.

6. The outcomes evaluated:

1. Mortality

  • In ICU
  • In Hospital

2. Infections

  • % of patients
  • Episodes of infections
  • Type of infection
    • Organ system involved
    • Primary endogenous (caused by micro-organisms found on throat or rectal admission screening swabs) / secondary endogenous (caused by micro-organisms not carried on admission screening throat or rectal swabs but acquired in throat or gut later in admission) / exogenous (caused by micro-organisms that were never present in throat or rectal screening swabs)
    • Micro-organisms involved
  • Timing of infection

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes, but no details of method of randomization have been included in the paper, so this makes the validity of randomization difficult to assess. 2. Were all patients who entered the trial properly accounted for and attributed at its conclusion? Was follow-up complete? Yes - n =107, initially enrolled 117 but 10 excluded because they didn't fulfil the inclusion criteria. All 107 were followed up and analysed
Were patients analyzed in the groups to which they were randomized? No. 117 patients were randomized but 10 were excluded after randomization (5 in each group).

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?


4. Were the groups similar at the start of the trial?

Mostly except for the Predicted ICU mortality which was higher in the control group 33.1% vs. 24.7% on the study group. However, it is not specified what scoring system was used for this but the authors have cited a study that describes infection classification, not a severity of illness score which may not give a true predicted ICU mortality score compared to those which are conventionally used.

  • Study group burnt total BSA 34%, control group 37.7%
  • Study group mean full thickness 19.3%, control group 19%
  • Study group inhalational injury present in 64%, control group 68%
  • Study group 74% were ventilated, control group 80
    • Demographically similar:

Study group 83% males, average age 41.4 years

Control group 74% males, average age 48.1years.

5. Aside from the experimental intervention, were the groups treated equally?
Yes - they all received standard care and any infections were treated with their unit protocol drugs then adjusted according to microbiological results. Standard care for the burn unit included resuscitation using the modified Parkland formula, enteral nutrition within 24 hours, excision and graft of burn within first 2 days, and closed dressings treated with silver sulfadiazine or iodine-povidone ointment.

III. What were the results?

1. How large was the treatment effect and how precise was the estimate of the treatment effect?

The table below summarizes all the statistics. For all types of pneumonia there is an absolute risk reduction (ARR) of 14.2%. However the 95% confidence intervals (CI) cross zero making the NNT range from a number needed to harm (the negative ARR) to a NNT (the positive end of the ARR's 95% CI) (see table).There does appear to be a large ARR for primary endogenous pneumonia of 31.5% within relatively tight 95% CI and a NNT of 3.2 with tight 95% CI.

There also appears to be a benefit to both ICU and hospital mortality with the SDD group, however the 95% CI for both the ARR and NNT are slightly wider so it is difficult to know how significant this really is. The same applies to the incidence of urinary tract infections.

Variable SDD incidence (%) Controls incidence (%) ARR % 95% CI for ARR NNT 95% CI for NNT RR
All types pneumonia 18/53(33.9)

26/54 (48.1)

14.2 -4.26 - +32.6 7 NNH 23 to NNT 3 0.7

Primary endogenous pneumonias

0/53 (0)

17/54 (31.5)


19.1 - 43.9

3.2 2 to 5 0

Secondary endogenous pneumonias

17/53 (32.1)

16/54 (29.6)


-19.9 - +15.1

NNH 40 NNH 5 to NNT 6 1.1

Exogenous pneumonias

5/53 (9.4) 3/54 (5.5) -3.9 -13.8-+6.1 NNH 25 NNH 7 to NNT 16 1.7

ICU mortality

5/53 (9.4)

15/54 (27.8)

18.3 4.0-32.6 5.5 3 to 25 0.34

Hospital mortality

6/53 (11.3)

15/54 (27.8)

16.5 1.8-31.1 6.1 3 to 55 0.4

Total infections

25/53 (43) 40/54 (74) 26.9 9.1-44.7 3.7 2 to 11 0.58

19/53 (35.8)

17/54 (31.5)

-4.4 -22.3%-+13.5 NNH 22 NNH 4 to NNT 7 1.1

Burn wound infection

10/53 (18.9)

11/54 (20.4)

15.0 -13.5-+16.5 7 NNH 7 to NNT 6 0.93

Urinary tract infection

6/53 (11.3)

14/54 (25.9)

15.1 0.5-29.7 6.6 3 to 200 0.44


Other results: No difference in days of mechanical ventilation of survivors
No difference in length of ICU stay of survivors
No difference in hospital length of stay of survivors

2. How precise was the estimate of the treatment effect?

The only result with real precision was the reduction in primary endogenous pneumonias, with a NNT of 3, with 95% CI of 2-5. Although not terribly precise, the 95% CI for ICU mortality reduction ranges from 3 - incredibly effective - to 25. Treating 25 patients with this methodology to prevent one death may be worthwhile.

IV. Will the results help me in caring for my patients?

  1. Can the results be applied to my patient care?

    • The only children involved were 14 years and above and there is no mention of how many children (i.e. ages 14 to 16 years) were included.
    • There was no age stratification. Extremes of age are important prognosticators for severe burns and even if this was taken into account in the scoring system for predicted ICU mortality it would be useful to see mortality figures for different age groups (1,2,3).
    • Only 80% of the patients were mechanically ventilated- this by definition would put the non-ventilated patients in a lower risk category for mortality. It may have been better to exclude all patients who did not require mechanical ventilation.
    • The reason for delayed admission (days 2 and 3 post injury) to the burns ICU is not specified. This makes it difficult to know whether these patients were stabilised in a non-burns ICU then transferred on day 2 or 3 to the burns ICU or whether they were well with no need for ICU initially then deteriorated requiring burns ICU admission. This will obviously change the risk of mortality. It may have been better to have included this data or to have used more strict inclusion criteria so that only those requiring immediate burns intensive care were included.
    • A trial involving children of all ages would be necessary before definite benefit could be assumed for children; however it is likely, based on the magnitude of the results that some benefit would be obtained both for reduction in mortality risk and reduction in infection risk. If such a trial was conducted including only the more severely burned patients (i.e. greater than 40% total BSA and, or inhalation injury) and including only those who require mechanical ventilation on the first day post-injury would give more useful results (2,3).
    • The regime used by the authors is also not strictly "selective digestive decontamination" as it utilizes a broad spectrum intravenous antibiotic.
  2. Were all clinically important outcomes considered? Yes, both mortality and infection were outcomes. Length of ICU and hospital stay was mentioned but the study was not powered to look at this.
  3. Are the likely treatment benefits worth the potential harms and costs?

    • Cost has not been mentioned but an NNT = 5 for mortality and NNT = 4 for all causes of infection is probably worthwhile from a cost point of view.
    • Potential harms are the apparent increased incidence of MRSA and pseudomonas infection in the treatment group; however MRSA was endemic in the ICU at the time so this may not be applicable to all ICUs. The authors have subsequently added vancomycin to the SDD regime that they have continued to use after the trial. An audit of the incidence of bacterial infections would be wise after this regime change to monitor the spread of potentially resistant organisms such as MRSA and VRE and pseudomonas.

References (if any)

  1. 1. Pereira CT et al. Age-dependent differences in survival after severe burns: a unicentric review of 1674 patients and 179 autopsies over 15 years. J Am Coll Surg. 2006 Mar;202(3):536-48 [abstract]
  2. 2. Gomez-Cia T et al. Mortality according to age and burned body surface in the Virgen del Rocio University Hospital. Burns. 1999 Jun;25(4):317-23. [abstract]
  3. 3. Brusselaers N et al. Outcome and changes over time in survival following severe burns from 1985 to 2004. Intensive Care Med (2005) 31:1648-1653 [abstract] founded 1995

Questions or comments?

Provide Feedback | Editorial Notes | About Us