Table 1. Primary studies and systematic reviews evaluating chlorhexidine-containing antiseptics for the prevention of blood culture contamination.
| Referencea | Study design | Antiseptics comparedb | Main outcomesc | Comments | Attributiond |
| Mimoz et al. 1999 [16] (M, C) | RCT | A: CHG 0.5% + ALC (?%); B: PVI aq 10% | A: 14/1019; B: 34/1022; p<0.05 | Advantage of CHG + ALC over PVI aq | Incorrect |
| Trautner et al. 2002 [17] (M, C) | RCTe | A: CHG 2% + IPA 70%; B: IPA 70% seq IT (I2 2%, ETH 47%) | A: 1/215; B: 3/215; NS | Study design equivalent to RCT | Correct |
| Barenfanger et al. 2004 [18] | Non-RCTf | A: CHG 2% + IPA 70%; B: IT (composition?) | A: 158/5802; B: 186/5936; NS | Composition of IT could not be clarified | Incorrect |
| Madeo et al. 2008 [19] | Non-RCTf | A: CHG 2% + IPA 70%; B: Unknown | A: 40/1870; B: 304/4072; p<0.05 | Weak study design, comparator unknown | Correct |
| McLellan et al. 2008 [20] | Non-RCTf | A: CHG 2% + IPA 70%; B: IPA 70% | Complex outcomesg | Weak study design, thoughtful analysis | Correct |
| Stonecypher 2008 [21] | Non-RCTf | A: CHG 2% + IPA 70%; B: PVI aq 10% | A: 23/687; B: 37/612; p<0.05 | Alcohol in arm A only revealed by correspondence | Incorrect |
| Suwanpimolkul et al. 2008 [22] (C) | RCT | A: CHG 0.5% + ETH 70%; B: PVI aq 10% | A: 34/1068; B: 74/1078; p<0.05 | Advantage of CHG + ALC over PVI aq | Correct |
| Tepus et al. 2008 [23] | Non-RCTf | A: CHG 2% + IPA 70%; B: IPA 70% seq IT (I2 2%, ETH 47%) | A: 169/7606; B: 251/7158; p<0.05 | Confounder: staff training before CHG + IPA study arm | Intermediate |
| Marlowe et al. 2010 [11] | Non-RCTf | A: CHG 3.15% + IPA 70%; B: PVI aq 10% | A: 72/4274; B: 122/4942; p<0.05 | Attribution criticised in letter to the editor | Incorrect |
| Washer et al. 2010 [24] | Cluster-randomised cross-over trial | A: CHG 2% + IPA 70%; B: IPA 70% seq PVI aq 10% ; C: IPA 70% seq IT (I2 2%, ETH 50%) | A: 41/4347; B: 25/4261; C: 32/4198; all NS | Use of IPA before PVI and IT in arms B and C, clarified by author | Correct |
| Malani et al. 2007 [25] | Systematic review | 4 eligible trials, 2 with CHG-containing antiseptics | No clear evidence; possible benefits from packaged kits and alcohol-based antiseptics | Results overall inconclusive | Correct |
| Caldeira et al. 2011 [26] | Systematic review | 6 eligible trials, 3 with CHG-containing antiseptics | Alcoholic products > non-alcoholic ones; ALC + CHG > PVI aq; CHG compounds vs iodine compounds inconclusive; ALC alone not inferior to iodine products | Article appropriately analyses different ingredients and compositions of antiseptics | Correct |
ALC, alcohol (when alcohol type not known); aq, aqueous; CHG, chlorhexidine gluconate; ETH, ethanol; IPA, isopropanol; IT, iodine tincture; PVI, povidone iodine; RCT, randomised clinical trial; seq, sequential application; vs, versus; ?%, percentage not specified; > (in systematic reviews), performing better than.
Annotation with (M) or (C) denotes whether original studies were included in the systematic reviews of Malani et al [25] (M) or Caldeira et al [26] (C).
A, B, and C denote different study arms.
Outcome: number of contaminated blood cultures per cultures obtained in each study arm. Significance is indicated either by NS (not significant) or p<0.05 (when significant).
Attribution: assesses whether study outcomes derived from alcohol plus CHG were attributed to CHG alone by authors.
In this trial, all subjects received both antiseptics at the same time, outcomes were assessed blindly.
These studies were classified as non-randomised cluster cross-over trials. Some had been conducted by prospective sequential implementation of different antiseptic regimens in clinical units [18], [20], [21], some by retrospective comparison of antiseptic regimens [11], [19], [23].
This study had complex outcomes from several pre- and post-intervention intervals showing that rigorous training and application may be more important than the choice of antiseptic.