Table 3. Primary studies and systematic reviews evaluating chlorhexidine-containing antiseptics for the prevention of surgical site infections.
| Referencea | Study design | Antiseptics comparedb | Main outcomesc | Comments | Attributiond |
| Berry et al. 1982 [50] (E, L, N) | RCT; mixed surgery, including abdominal | A: CHG 0.5% + ALC (?%); B: PVI 10% + ALC (?%) | A: 44/453; B: 61/413; p<0.05 | ALC type and content in both study arms unknown; difference significant | Incorrect |
| Brown et al. 1984 [51] (L, N) | RCT; mixed surgery, including obstetric, abdominal | A: CHG 0.5% + IPA 70%; B: PVI aq (0.7% av I2) seq PVI aq (?%) | A: 23/378; B: 29/359; NS | Difference non-significant | Incorrect |
| Ostrander et al. 2005 [52] (L) | RCT; clean foot and ankle surgery | A: CHG 2% + IPA 70%; B: IPOV (0.7% av I2) + IPA 74%; C: Chloroxylenol 3% | A: 1/40; B: 0/40; C: 2/40; all NS | Also skin microbial counts studied, but methodology not adequately described | Intermediate |
| Veiga et al. 2008 [53] (L) | RCT; elective clean plastic surgery | A: CHG 0.5% + ALC (?%); B: PVI 10% + ALC (?%) | A: 0/125; B: 4/125; NS | Difference non-significant; ALC type and content unknown | Incorrect |
| Cheng et al. 2009 [54] | RCT; clean forefoot surgery | A: CHG 2% + IPA 70%; B: PVI 10% + IPA 23% | A: 0/25; B: 0/25; NS | Small study; focus on skin counts; IPA content in arm B far below active range | Intermediate |
| Paocharoen et al. 2009 [55] (L, N) | RCT; general surgery, including clean, clean-contaminated and contaminated cases | A: CHG 4% + IPA 70%; B: PVI aq (?%) | A: 5/250; B: 8/250; NS | Difference non-significant | Incorrect |
| Saltzman et al. 2009 [56] (L) | RCT; clean shoulder surgery, including arthroscopic | A: CHG 2% + IPA 70%; B: IPOV (0.7% av I2) + IPA 74%; C: PVI aq scrub & paint (0.75% & 1.0% av I2) | A: 0/50; B: 0/50; C: 0/50; NS | Small study; focus on skin counts; microbiological methods potentially inadequate | Correct |
| Swenson et al. 2009 [57] (N) | Non-RCT; mixed general surgerye | A: CHG 2% + IPA 70%; B: PVI aq 7.5% seq IPA 70% seq PVI aq 10%; C: IPOV (0.7% av I2) + IPA 74% | A: 68/827; B: 72/1514; C: 38/794; A:B, A:C p<0.05 | Significantly more infections in CHG + ALC arm, but only superficial ones | Correct |
| Darouiche et al. 2010 [6] (L, N) | RCT; mixed clean-contaminated surgery, including abdominal | A: CHG 2% + IPA 70%; B: PVI aq 10% scrub & paint | A: 39/409; B: 71/440; p<0.05 | Seminal study; significant difference in favour of CHG + ALC over PVI aq | Correct |
| Sistla et al. 2010 [58] | RCT; elective clean inguinal hernia surgery | A: CHG 2.5% + ETH 70%; B: PVI aq 10% | A: 14/200; B: 19/200; NS | Difference non-significant | Correct |
| Levin et al. 2011 [59] | Non-RCT; elective gynaecological laparotomy surgerye | A: CHG aq 2% seq IPA 70%; B: PVI aq 10% seq PVI 10% + ETH 65% | A: 5/111; B: 21/145; p<0.05 | Weak study design; significant difference | Correct |
| Edwards et al. 2004 [60] | Systematic review | 7 eligible trials, only 1 with a CHG-containing arm [50] | Overall inconclusive due to lack of well-designed studies | Review from 2004, updated 2009; lack of studies at the time | Intermediate |
| Lee et al. 2010 [12] | Systematic review | 9 eligible trials, 5 studied CHG + ALC vs PVI aq, 4 studied CHG + ALC vs PVI + ALC (including 1 both), 1 studied CHG aq vs PVI aq for mucous membranes | “Chlorhexidine” superior to iodine, based on majority CHG + ALC vs PVI aq outcomes | Analysed both infection rates and microbial skin counts; criticised in letters to the editor | Incorrect |
| Noorani et al. 2010 [13] | Systematic review | 6 eligible trials, 3 studied CHG + ALC vs PVI aq, 2 CHG + ALC vs PVI + ALC, 1 CHG aq vs PVI aq for mucous membranes | “Chlorhexidine” superior to iodine, based on majority CHG + ALC vs PVI aq outcomes | Attribution criticised in letters to the editor | Incorrect |
ALC, alcohol (when alcohol type not known); aq, aqueous; av, available (referring to available iodine as opposed to total iodine complex); CHG, chlorhexidine gluconate; ETH, ethanol; IPA, isopropanol; IPOV, iodine povacrylex; PVI, povidone iodine; RCT, randomised clinical trial; seq, sequential application; vs, versus; ?%, percentage not specified.
a
Annotation with (E), (L), or (C) denotes whether original studies were included in the systematic reviews of Edwards et al [60] (E), Lee et al [12] (L), or Noorani et al [13] (N).
b
A, B, and C denote different study arms.
c
Outcome: surgical site infections per number of surgical procedures in each study arm. Significance is indicated either by NS (not significant) or p<0.05 (when significant).
d
Attribution: assesses whether study outcomes derived from alcohol plus CHG were attributed to CHG alone by authors.