ChEETAh: Routine Sterile Glove and Instrument Change at Abdominal Wound Closure

Bottom line Positive

ChEETAh randomized 81 hospitals in seven low- and middle-income countries to routine glove and instrument change before abdominal wound closure versus current practice; 30-day surgical site infection was 16.1% versus 18.9% (adjusted risk ratio 0.87, 95% CI 0.79–0.95).

Adjusted RR 0.87 for 30-day SSI; the design targeted a 4-point absolute reduction.

ChEETAh was a pragmatic, multicenter cluster-randomized trial that assigned 81 hospitals in seven low- and middle-income countries to routine change of sterile gloves and instruments before abdominal wound closure or to current practice, in 13,301 adults and children having clean-contaminated, contaminated or dirty abdominal surgery. Surgical site infection (SSI) within 30 days occurred in 931/5,789 patients (16.1%) with the change and 1,280/6,768 (18.9%) without it (adjusted risk ratio 0.87, 95% CI 0.79–0.95; P = 0.0032).^[1]

Design and population

Pragmatic multicenter cluster-randomized superiority trial, hospitals randomized. Adults and children having emergency or elective clean-contaminated, contaminated or dirty abdominal surgery through an incision of at least 5 cm, excluding cesarean section, in 81 hospitals in seven low- and middle-income countries.

Arms

Routine change of sterile gloves and instruments before wound closure versus current hospital practice

Report and follow-up

2022 · 30 days after surgery; primary outcome analyzed in 12,557 of 13,301 patients

Endpoint	Results	Effect	Finding
SSI within 30 days (superficial or deep incisional, CDC criteria); adjusted risk ratio	931/5,789 (16.1%) vs 1,280/6,768 (18.9%)	RR 0.87; CI 0.79–0.95; P 0.0032	Met

Harms: Serious adverse events: The authors did not anticipate or collect specific data on serious adverse events.; Mortality within 30 days: 394/6,110 (6.4%) vs 455/7,095 (6.4%); Unexpected readmission for a wound-related problem: 194/5,761 (3.4%) vs 243/6,675 (3.6%)

NCT03980652 · Report DOI

Design

The trial ran in Benin, Ghana, India, Mexico, Nigeria, Rwanda and South Africa between June 24, 2020 and March 31, 2022. Any hospital doing non-cesarean elective or emergency abdominal surgery could join, and hospitals with an existing routine glove or instrument change policy were excluded (none were identified). Patients of any age were eligible for abdominal surgery for any indication, including trauma, when the operation was clean-contaminated, contaminated or dirty and the abdominal incision was at least 5 cm. Clean surgery and cesarean delivery were excluded. Guidelines from WHO (2018), the US Centers for Disease Control (2017) and NICE (2019) made no recommendation on routine change of gloves and instruments before the trial, citing a weak evidence base.^[1]

Hospitals were randomized 1:1 centrally at the Birmingham Clinical Trials Unit with minimization by country and referral status, and learned their allocation before site training. Only intervention-group teams were trained. Because the allocation was at hospital level, patients gave no consent for exposure to the intervention, only for 30-day outcome assessment. Site investigators and outcome assessors could not be masked; patients did not know their hospital's allocation.^[1]

The primary outcome was superficial or deep incisional SSI within 30 days by CDC criteria. Assessors, in hospital, in the community or by telephone, answered a script of predefined questions mapped to the CDC criteria and did not make a binary SSI call. The analysis excluded patients who were lost to follow-up, refused consent, died before assessment (unless SSI had been recorded at discharge) or had missing data, and used a multilevel log-binomial model adjusted for hospital type, contamination, operative approach and timing of surgery. Secondary outcomes were 30-day mortality, SSI at discharge, unexpected readmission and reoperation for a wound-related problem, return to normal activities and length of stay, with no adjustment for multiple comparisons. The sample size assumed a 16% control rate and a 4-point absolute (25% relative) reduction, the smallest difference an international surgeon group considered clinically important.^[1]

Results

The primary outcome was analyzed in 12,557 of 13,301 patients. SSI occurred in 931/5,789 (16.1%) in the intervention group and 1,280/6,768 (18.9%) in the current-practice group (adjusted risk ratio 0.87, 95% CI 0.79–0.95; P = 0.0032), meeting the primary hypothesis.^[1]

The direction of effect held in the per-protocol analysis (919/5,693 [16.1%] versus 18.9%; risk ratio 0.84, 95% CI 0.76–0.93), in the best-case (0.81, 95% CI 0.72–0.92) and worst-case (0.86, 95% CI 0.78–0.95) scenarios for patients with missing outcome data, and among hospitals that reached the 200-patient recruitment target (684/4,127 [16.6%] versus 1,050/5,313 [19.8%]; 0.82, 95% CI 0.74–0.91). Adjusting for the minimization factors alone (country and hospital type) gave a risk ratio of 0.66 (95% CI 0.60–0.73). The authors reported no heterogeneity of effect across the prespecified subgroups (country, hospital type, urgency, contamination, midline versus non-midline incision, and child versus adult), which the trial was not powered to test.^[1]

No secondary outcome differed between groups. SSI at hospital discharge was 484/5,976 (8.1%) with the change versus 706/7,002 (10.1%) without (risk ratio 0.77, 95% CI 0.52–1.12), return to normal activities 63.9% versus 66.6% (0.99, 95% CI 0.83–1.17), and median length of stay 6 versus 5 days (geometric mean ratio 1.12, 95% CI 0.97–1.25).^[1]

A model-based cost-effectiveness analysis pooled the trial's outcomes from all seven countries with resource use and unit costs from India (a costing study within the FALCON trial). The average cost per surgery was $259.92 with the change and $261.10 without it (difference −$1.18, 95% CI −$4.08 to $1.33; 2021 US dollars, provider perspective), and the model estimated SSI in 17.6% versus 19.7%. Unit costs from Mexico and Ghana gave similar conclusions.^[2]

Surgery and treatment delivery

The operating surgeon, assistants and scrub staff all changed their sterile gloves (the outer gloves if double gloved) and used a separate sterile set containing a needle holder, forceps and scissors for abdominal wall closure. The change was to take place after the abdominal part of the operation and before the wound edges were handled. Hospitals set the instruments aside at the start of the case, for example wrapped on a clean swab, or opened a new instrument pack. Skin preparation, wound edge protectors, triclosan sutures, wound washout and negative pressure wound therapy were left to the surgeon in both groups, and control teams could change gloves or instruments if that was local practice.^[1]

The glove-and-instrument change took place in 6,044/6,144 patients (98.4%) in the intervention group; in the current-practice group, 58/7,157 (0.8%) had at least one component changed. Among the 5,789 intervention patients in the primary analysis, 100 did not receive the change; in the control group, 45 had a glove change only, 3 an instrument change only and 10 both. The number who had one component without the other was too small to analyze separately. Mölnlycke Healthcare donated sterile BioGel gloves to intervention hospitals.^[1]

Of the 13,301 patients, 11,825 (88.9%) were adults, 6,125 (46.0%) had elective surgery, 8,086 (60.8%) had a clean-contaminated operation and 5,215 (39.2%) a contaminated or dirty one. Nearly all operations were open (13,011, 97.8%), with midline incisions in 8,129 (61.1%). Urologic operations numbered 679 (293/6,144 [4.8%] intervention, 386/7,157 [5.4%] control), alongside gynecologic, colorectal, appendiceal, hepatopancreatobiliary and other abdominal operations. Prophylactic antibiotics were given within 60 minutes of incision in 98.3% of the intervention group and 98.4% of controls.^[1]

Adverse events

The investigators did not anticipate or collect specific data on serious adverse events. Thirty-day mortality was 394/6,110 (6.4%) with the change and 455/7,095 (6.4%) without it (risk ratio 0.88, 95% CI 0.70–1.10), unexpected readmission for a wound-related problem 194/5,761 (3.4%) versus 243/6,675 (3.6%) (1.02, 95% CI 0.75–1.39), and unexpected reoperation for a wound-related problem 89/5,760 (1.5%) versus 157/6,674 (2.4%) (0.73, 95% CI 0.48–1.10).^[1]

Limitations

Teams and assessors were unmasked, and most 30-day assessments were by telephone (80.4% intervention, 83.0% control of patients with follow-up data). The primary analysis excluded 355 intervention and 389 control patients, among them 313 and 309 who died before outcome assessment; best-case and worst-case analyses gave consistent results.^[1]

Case mix differed between groups, as expected when hospitals are randomized: emergency surgery was 3,050/6,144 (49.6%) with the change versus 4,126/7,157 (57.6%) without it, and gynecologic operations were 19.6% versus 26.1% while colon and rectal operations were 13.7% versus 9.0%. The estimate depends on this adjustment: the primary model gave a risk ratio of 0.87, and the model adjusted for minimization factors alone gave 0.66.^[1]

The control-group SSI rate (18.9%) was higher than the 16% assumed in the sample size, and the difference between groups was smaller than the 4 points the trial was sized to detect. The trial tested glove and instrument change together, did not record single versus double gloving, and included 679 urologic operations in a cohort in which 97.8% of operations were open.^[1]

In conversation

Same setting Darouiche (2010)

The investigators identified two earlier randomized trials: a single-center US bundle trial that included glove and instrument change (233 patients having colorectal, gynecologic or urologic oncologic operations) and a Japanese trial of instrument change for fascial closure (453 elective colorectal patients). Both were at moderate or high risk of bias and neither showed a significant SSI reduction. Other evaluations were non-randomized and tested glove and instrument change within bundles of up to 13 other measures, so ChEETAh tested it as a single intervention at a larger scale than those trials.^[1]

Guidelines and approvals

• Hands Deserve Better global consensus (J Hosp Infect, 2026). Statement 10: during general surgery procedures, surgical team members should change gloves before abdominal closure as part of a bundle . The supporting text describes ChEETAh. Mölnlycke Health Care funded the consensus and also funded ChEETAh.^[3]

Report history

ChEETAh report history

Published	Report	Cutoff	Headline results	Citation
2022-10-31	Primary report · primary	—	SSI within 30 days (superficial or deep incisional, CDC criteria); adjusted risk ratio: 931/5,789 (16.1%) vs 1,280/6,768 (18.9%); RR 0.87; CI 0.79–0.95; P 0.0032 Met Per-protocol sensitivity analysis; adjusted risk ratio: 919/5,693 (16.1%) vs 1,280/6,768 (18.9%); RR 0.84; CI 0.76–0.93; P 0.0010 Best-case scenario as labeled in Figure 3 (patients with missing primary-outcome data recoded); adjusted risk ratio: 931/5,831 (16.0%) vs 1,280/6,848 (18.7%); RR 0.81; CI 0.72–0.92; P 0.0009 Worst-case scenario as labeled in Figure 3 (patients with missing primary-outcome data recoded); adjusted risk ratio: 973/5,831 (16.7%) vs 1,360/6,848 (19.9%); RR 0.86; CI 0.78–0.95; P 0.0059 Only hospitals that reached the 200-patient recruitment target; adjusted risk ratio: 684/4,127 (16.6%) vs 1,050/5,313 (19.8%); RR 0.82; CI 0.74–0.91; P 0.0004 Adjusted for minimization factors only (country and hospital type); adjusted risk ratio: 931/5,789 (16.1%) vs 1,280/6,768 (18.9%); RR 0.66; CI 0.6–0.73; P <0.0001 SSI at discharge; adjusted risk ratio: 484/5,976 (8.1%) vs 706/7,002 (10.1%); RR 0.77; CI 0.52–1.12; P 0.17 Self-reported return to normal activities by 30 days; adjusted risk ratio: 3,651/5,717 (63.9%) vs 4,413/6,623 (66.6%); RR 0.99; CI 0.83–1.17; P 0.89 Median (IQR) days; geometric mean ratio: 6 (4–9) vs 5 (3–8); —	NIHR Global Research Health Unit on Global Surgery. Routine sterile glove and instrument change at the time of abdominal wound closure to prevent surgical site infection (ChEETAh): a pragmatic, cluster-randomised trial in seven low-income and middle-income countries. *Lancet.* 2022;400(10365):1767-1776. doi:10.1016/S0140-6736(22)01884-0 DOI
2024	Cost-effectiveness analysis · secondary	—	Average cost per surgery, base case (India unit costs, 2021 US$, provider perspective); mean difference, intervention minus current practice: $259.92 vs $261.10; difference -1.18; 95% -4.08–1.33 Model-estimated percentage of patients with SSI, base case; absolute risk reduction in percentage points: 17.6% vs 19.7%; difference 2.1; 95% 2.07–2.84	NIHR Global Health Research Unit on Global Surgery. Routine sterile glove and instrument change at the time of abdominal wound closure to prevent surgical site infection (ChEETAh): a model-based cost-effectiveness analysis of a pragmatic, cluster-randomised trial in seven low-income and middle-income countries. *Lancet Glob Health.* 2024;12(2):e235-e242. doi:10.1016/S2214-109X(23)00538-7 DOI

---

References

1. NIHR Global Research Health Unit on Global Surgery. "Routine sterile glove and instrument change at the time of abdominal wound closure to prevent surgical site infection (ChEETAh): a pragmatic, cluster-randomised trial in seven low-income and middle-income countries." Lancet. 2022;400(10365):1767-1776. doi:10.1016/S0140-6736(22)01884-0

2. NIHR Global Health Research Unit on Global Surgery. "Routine sterile glove and instrument change at the time of abdominal wound closure to prevent surgical site infection (ChEETAh): a model-based cost-effectiveness analysis of a pragmatic, cluster-randomised trial in seven low-income and middle-income countries." Lancet Glob Health. 2024;12(2):e235-e242. doi:10.1016/S2214-109X(23)00538-7

3. Enz A, Boermeester MA, Chatterjee A, et al. "Hands Deserve Better: global clinical consensus recommendations on surgical gloving practice." J Hosp Infect. 2026;172:115-128. doi:10.1016/j.jhin.2026.03.025