Double gloving for self-protection in high-risk surgeries: a systematic review and meta-analysis

Abstract

Background

Double gloving is recommended for protecting surgical personnel from infections, but it is not a universal practice, especially in low- and middle-income countries where risk is very high. Evidence for double gloving is still only moderate, and for indicator double gloves, it is even rare. This systematic review and meta-analysis includes recent trials to analyse outcomes like glove perforations (inner/outer/matched/intraoperatively detected) and hand contamination rates for single versus double including indicator double-gloved conditions and identify factors to be considered for deciding double gloving.

Method

Six databases PubMed, EBESCO, Embase, CINAHL, Scopus, Web of Science, and CENTRAL were searched up to May 2024. The quality of included trials was assessed using Cochrane risk-of-bias tool (version 5.1.0). Heterogeneity among trials was estimated using the chi-squared (I²) test. RevMan 5.3 was used for meta-analysis and subgroup analysis. Odds ratio at 95% confidence interval was used as statistical measure to compare outcomes and calculate effect size. Publication bias was assessed through a funnel plot.

Result

A review of these total of 18 randomized controlled trials showed that deep/major/emergent surgeries, primary surgeons, and longer surgical duration are prone to have higher glove perforations. Impaired dexterity is not a constraint for double gloving and has no impact on glove perforations. Meta-analysis of outcomes suggests that double gloving (standard or indicator) provides significant protection against infections compared to single gloves in terms of reduced inner (OR = 0.2, 95% CI 0.14–0.31) and matched glove perforations (OR = 0.1, 95% Cl 0.07–0.13) and lower incidences of hand contamination (OR = 0.28, 95% Cl 0.14–0.54). Standard double gloves were more effective in reducing matched glove perforations than indicator double gloves. But for detecting glove perforations intraoperatively, only the indicator double glove (OR = 8.64, 95% Cl 4.78–15.61) was effective.

Conclusion

Double gloving is recommended over single gloving for better safety of surgical personnel and indicator gloves for better detection of perforations during surgery so that it can be changed timely, but it does not provide any additional protection. In the future, there should be high-quality trials for specific surgeries, surgical personnel, and different surgical durations taking into consideration the cost-effectiveness of indicator gloving over standard double gloving so that specific recommendations can be made.

Introduction

Surgical personnel are at high risk of contracting infections during invasive surgeries. Risk is even higher in low- and middle-income countries (LMICs) where the patient population has a prevalence of HIV, or hepatitis B/C (HBV/HCV), and when there is extensive exposure to the patient’s blood and body fluid [1]. As per the World Health Organization (WHO) estimate, globally, about 83,000 healthcare workers contract bloodborne infections annually due to percutaneous injuries during surgeries and patient care of which 10–65% were reported from LMICs [2]. The compromised health of surgical personnel due to occupational exposure to infection not only affects them individually but also affects tremendously overall healthcare quality [3].

Many studies including Cochrane intervention reviews recommended double gloving as one of the strategies to increase protection [4,5,6,7,8,9,10,11]. Wearing indicator gloves is suggested to be further protective by detecting perforations intraoperatively [8, 12]. However, double gloving is still not a standard practice in the UK, Europe, and the USA [12], despite the recommendations of many organizations like WHO [13], Centers for Disease Control and Prevention [14], and the American Operating Nurses Association [15]. Even the most recent meta-analysis suggested the quality of evidence for double gloving as moderate [16]. Data on the usefulness of double gloving over single gloving in resource-constrained settings of LMICs is very limited. Further, it is not clear if indicator gloves provide extra protection over standard double gloves. Many factors influence the decision to double gloving or glove perforations like type of surgery and surgical personnel, duration of surgery, and dexterity [9, 17] that need to be reviewed systematically.

So, this systematic review and meta-analysis includes recent trials from LMICs and comprehensively compares different outcomes (glove perforations, hand contamination, intraoperative detection of perforations) for single and double gloving including indicator gloving and identified factors that could be important for deciding appropriate gloving practice. We also analyzed the efficacy of double gloving for factors like surgical duration, the economic status of the country, and the year of the trial to see if these can impact the outcomes. This meta-analysis will provide updated and better evidence for double gloving including indicator double-gloving practices in different surgical environments and also identify the areas to be focused on so that more specific recommendations can be made.

Method

This review was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines.

Search method for identification of studies

The relevant articles for this review were searched systematically by browsing different databases that include PubMed, MEDLINE, Embase, CINAHL, Scopus, Web of Science, and CENTRAL. For devising the primary search strategy, study questions were divided into population, intervention, and outcome (PIO). For each, their MeSH terms, subject headings or major headings, various entry terms, and synonyms were combined using OR as a Boolean operator, and then the search strings were combined using the AND Boolean operator to get the final result. The period taken was from the inception of the database to May 2024, and the language was limited to English only. The search strategy was optimized for different databases, and its details are provided in Appendix I. Apart from databases, reference lists of published systematic reviews and articles were also checked. All the search results were then combined and after removing duplicates screened for eligibility. Additionally, to reduce publication bias, the System for Information on Grey Literature was also searched.

Eligibility criteria for studies

Two authors independently screened the searched articles systematically by first going through the title, then the abstract, and finally the full text. The following inclusion criteria were applied.

• The study should be a randomized controlled trial (RCT).

• Involve surgical personnel including surgeons, assistant surgeons, or scrub nurses.

• Have compared standard or indicator double gloves to single gloves

• Have reported at least one of the following outcomes: glove perforations (single/inner/outer/matched/intraoperatively detected) and incidences of blood/body fluid contamination including needle stick injuries

• Have used at least one of two methods of perforation detection, i.e. water leak¹ or air inflation tests²

Exclusion criteria

Articles that are not RCTs (observational studies, animal studies, or studies performed in simulated surgical environment), were not in English, whose full text cannot be assessed, and not peer-reviewed, case reports, editorials, letters to the editors, opinion articles, and conference abstracts were excluded from the study. Studies, where the intervention was other gloving patterns and gloves are not of latex material, were also excluded.

Data collection and analysis

Authors based on inclusion and exclusion criteria selected the studies to be included and obtained the full text. Two authors independently extracted the data from the included studies in an Excel spreadsheet following the checklist based on the Cochrane Handbook for Systematic Reviews of Interventions [18], and any conflict was resolved by consulting the corresponding author. Data extracted are study details (author, year, country, study participants, type of surgery, surgical duration, glove type, method of testing perforation), intervention (total number of single and double gloves including standard and indicator), and outcomes (single/inner/outer/matched/intraoperatively detected perforations, incidences of blood/body fluid contamination including needle stick injuries) for meta-analysis along with other outcomes reported like most common site of perforation, rate of perforation for surgical personnel involved, and type and duration of surgery.

The unit of measurement taken in this study for outcomes, single/inner/outer/matched perforations, and hand contamination was the number of gloves perforated/incidences of contamination out of the total number of single or double gloves (outer or inner). The incidences of hand contamination were considered as any injury or needle stick injury or any contamination of the skin with blood or body fluid noticed or reported during surgery. Matched perforations are those when both inner and outer gloves are perforated at a similar location in a double glove. For matched perforation and hand contamination rates, the total number of inner gloves is taken as the denominator. The outcome, intraoperative detection of perforations, was measured as the total number of perforated gloves detected during surgery out of the total perforated single or double gloves. The unit of measurement for outcomes was converted into the desired unit based on the information provided in the article wherever necessary.

Quality assessment

The quality of the included trials was critically appraised independently by two authors using the Cochrane Collaboration tool to assess the risk of bias [18]. It evaluates the study design and the integrity and objectivity of data in terms of selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases and classifies each as high risk, unclear risk, or low risk of bias. If the glove perforation was tested using only one of the two methods, then it is considered a potential source of bias (other bias). Results were entered in RevMan 5.3 after resolving the conflicting assessment through discussion with the corresponding author.

To assess the overall risk of bias in studies, random sequence generation, allocation concealment, and blinding of the outcome assessor are considered the most important aspects. Studies that have a high risk in any of these three are considered to have an overall higher risk of bias.

Statistical analysis

Data from studies reporting the same outcome were pooled, and effect size was calculated using RevMan 5.3 software. The odds ratio at a 95% confidence interval (Cl) was used as a statistical measure to compare outcomes and calculate effect size. Heterogeneity among studies included was assessed using a chi-squared test (I²) and rated as unimportant (0–40%), low (30–60%), moderate (50–90%), and high heterogeneity (75–100%) as described in the Cochrane Handbook for Systematic Reviews of Interventions [18]. The heterogeneity of more than 50% was considered as a cutoff to apply a random effect model else a fixed effect model (I² < 50%) was used [18]. Publication bias was assessed using a funnel plot and with Egger’s test for outcomes that had more than 10 studies included [18]. Subgroup analysis was done to assess the effect of indicator gloves, economic status of countries, year of study, and duration of surgery on the outcome. A sensitivity analysis was done for all the outcomes which showed positive results by reanalysing the data excluding trials that have an overall high risk of bias. A narrative summary of all the included studies was also made which includes study details.

Results

Search results

A total of 729 articles were retrieved, out of which 531 were excluded after screening through the title and 168 after screening through the abstract for not being relevant based on exclusion criteria. The full text of 32 articles was reviewed out of which 13 were excluded for not being RCTs, 1 article was excluded for a unit of measurement being perforations per operation and not per glove [19], and 1 other for only visually assessing the perforations [20]. A total of 18 trials were included in the meta-analysis (Fig. 1).

Fig. 1

The Preferred Reporting Items for Systematic review and Meta-Analysis (PRISMA) flow diagram indicating the selection of relevant studies

Characteristics of included studies

The 18 RCTs [4,5,6,7, 21,22,23,24,25,26,27,28,29,30,31,32,33,34] were included in this meta-analysis, 11 [4, 6, 21, 24,25,26,27, 29, 30, 32, 34] were carried out in developed (UK, Spain, Australia, France, Finland, Denmark) and 7 [5, 7, 22, 23, 28, 31, 33] in developing (India, Hongkong, Thailand, Nigeria) countries between 1999 and 2023 (10 studies: 1990–2000; 8 studies: 2001–2023). Study periods across different studies ranged between 1 and 20 months. Four studies undertook large trials with more than 1000 gloves [21, 22, 28, 30], 9 have a trial size of 500–1000 gloves [6, 7, 23,24,25,26,27, 33, 34], while 5 have less than 500 gloves analysed [4, 5, 29, 31, 32] during the study. Prior sample size calculation was done in only two trials [4, 32]. Eleven studies used standard double gloves [4, 5, 7, 22, 23, 26, 28, 31,32,33,34], 3 studies used indicator gloves [21, 24, 29], while 3 used both a double indicator and standard double gloves separately in a two-arm study [6, 27, 30] as a counter-intervention to single gloves. One study used a mix of double-indicator and double-standard gloves [25] and considered into indicator double glove category. Seven trials also used unused pairs of gloves as blank control [7, 22, 25, 28, 31, 33, 34]. All the gloves used are made of latex material. The number of surgeries included in each trial varied between 66 and 885 with a total of 4450 and an average of 247 surgeries per trial. The majority of trials (9) were conducted in abdominal surgeries like caesarean section, hysterectomy, colectomy, hernia repair, hepatectomy, myomectomy, gastrointestinal surgery, obstetrics, and gynaecology surgery [4,5,6, 21,22,23,24, 32, 33]. Five trials were conducted in a combination of various types of surgeries, including general surgery, urologic surgery, plastic surgery, paediatrics surgery, vascular surgery, skin surgeries, thoracic surgeries and abdominal surgeries, neurosurgery, and orthopaedic and trauma surgery [7, 26, 28, 30, 31]. One trial each was conducted in plastic and reconstructive surgery [34], orthopaedic and trauma surgery [27], dental surgery [25], and vascular surgery [29]. Out of these few studies also categorized surgeries into major/deep versus minor/superficial, open versus laparoscopic, and elective versus emergency. Four studies [7, 21, 24, 28] reported higher glove perforation rates for major/deep surgeries than minor/superficial surgeries out of which in three trials [7, 21, 24] the difference was significant (p < 0.05). Four studies [7, 27, 28, 30] reported higher glove perforation in emergent surgeries than elective ones, and the difference was significant in two studies [27, 30]. Glove perforations are found to be higher in open surgeries than laparoscopic ones as reported in three studies [6, 24, 27] with the difference being significant in two trials [6, 27]. Surgical personnel involved in the included studies are primary surgeons, assistant surgeons, and scrub nurses. Five studies have all three involved in the surgery [7, 21, 26, 28, 34], and 11 studies have a primary surgeon and/or assistant surgeon involved [5, 6, 22,23,24, 27, 29,30,31,32,33], while in 1 study only scrub nurses participated in the trial [4]. In another study, dental surgeons and dental hygienists were the study subjects [25]. Eleven trials compared glove perforation rates for surgical personnel involved, out of which 9 reported primary surgeons having the highest glove perforation rate, and the difference was significant in 7 trials [6, 7, 21, 25, 26, 30, 32] and nonsignificant in two studies [27, 31]. Assistant surgeons [28] had significantly higher rates of glove perforation in one study, while another study [34] reported scrub nurses sustained maximum glove perforations among other surgical personnel. Few of the included studies gave information on the duration of surgery which ranged between less than 30 min and more than 180 min. Few studies compared glove perforation rates for the duration of surgery, and the majority of them showed that surgical procedures lasting more than 2 h (six trials) or even 1 h (one trial) had significantly higher glove perforation rates than those lasting less than 2 h [7, 24, 28, 30, 34] or 1 h [27]. Gani et al. (1990) also reported an increase in glove perforation rates for the duration of surgery [26]. However, three other trials did not observe any significant correlation between the duration of surgery and glove perforation rates [6, 31, 32]. Thirteen trials reported the most perforated location of the hand as one of the outcomes, which in 8 trials was found to be the index finger of the nondominant hand [6, 7, 23, 26,27,28, 30, 33]. Two studies reported middle fingers of nondominant hands [4, 22], one study reported middle and index fingers equally [5], and one study thumb [34] as the most perforated location, while one study only stated nondominant hand [21] and did not specify the finger. Seven trials reported data on the satisfaction of surgical personnel with dexterity when double-gloved. In five trials, the majority (60–100%) of surgical personnel reported feeling comfortable and satisfied [7, 22, 23, 31, 33], while in two trials a significant number of surgical personnel reported impaired dexterity when double-gloved than single-gloved [25, 32] (Table 1).

Table 1 Study details of the included trials

Quality assessment of included studies

Among 18 included trials, in 13 trials, surgical personnel were used as a unit of randomization [4,5,6, 21,22,23, 25, 27, 28, 31,32,33,34], 4 trials randomized patients [7, 26, 29, 30], while 1 trial randomized surgical procedure [24]. Ten trials were assessed to have an unclear risk of selection bias as the method of randomization was not clearly defined [5,6,7, 21, 24, 25, 27, 28, 31, 34]. Five trials considered having a low risk of selection bias as three studies used envelopes containing random numbers representing each group [22, 23, 33], one study used a random number table [32], while another used the Excel method [4] for randomization. Three trials have a high risk of selection bias as two trials used the date of birth of patients [29, 30] and one study used patient’s hospital record numbers [26] for randomization. Eleven trials were considered to have an unclear risk of bias [4,5,6, 21, 22, 24, 25, 27, 28, 33, 34] due to allocation concealment (selection bias) as no sufficient information was provided. Four trials had a low risk of bias [7, 23, 31, 32] as they used sealed envelopes for allocation concealment. Three trials had a high risk of bias as they used inadequate concealment, i.e. patient date of birth [29, 30] and hospital record number [26]. Since blinding surgical personnel to gloving patterns is practically not possible and it is considered not to affect the outcome, so all the trials were assessed as having a low risk of performance bias. The risk of bias from blinding of the outcome assessor (detection bias) was considered unclear for 10 trials, as no information was provided in 8 trials [6, 7, 21, 25, 27,28,29, 31], while in 1 trial [32] gloves were put in bags with the detailed questionnaire and in another trial [30] gloves were collected in labelled bags, and it is unclear if the experimenter has access to it. Six trials had a high risk of detection bias, as in two trials each glove was labelled with details [24, 34], in three trials gloves were collected in labelled bags assessed by the investigator [4, 22, 23], while in one trial surgical personnel themselves tested for the perforation [26]. Only one trial has a low risk of detection bias as surgeons put gloves in numbered bags and recorded details in a record sheet to which the experimenter was kept blind [5]. The risk of attrition bias was assessed as low for 8 studies, as 6 of them included all the gloves including changed ones in the study [4, 24, 25, 30], 1 trial excluded 12 gloves due to parenteral injuries [34], 1 excluded 6 gloves which are not properly labelled [32], while another excluded gloves from those surgeons who do not immediately replace the perforated gloves [26]. Seven studies considered having an unclear risk of attrition bias as no sufficient information was provided on which gloves are included or excluded [6, 21,22,23, 27, 28, 31]. Three trials have a high risk of attrition bias as they included only the first or original pair of gloves in the analysis and no information was provided on how many gloves were excluded [5, 7, 33]. All the included trials have a low risk of reporting bias as they reported all the outcomes. Three trials used both water leak and air inflation methods so considered having a low risk of other bias [4, 28, 31]. Fifteen trials have a high risk of other bias out of which 12 used only water leak tests [6, 7, 21, 24,25,26,27, 29, 30, 32,33,34] and 3 trials used only the air inflation test [5, 22, 23] (Figs. 2 and 3).

Fig. 2

Risk-of-bias categorization assessed by the authors for each domain and each included study

Fig. 3

Cumulative risk-of-bias graph of all the included trials

Meta-analysis

Out of the total 18 trials included in this meta-analysis, all reported glove perforation as primary outcomes, and few of them also reported blood contamination or needlestick injuries as secondary outcomes. All the outcomes (inner/outer/matched glove perforations, incidences of hand contamination, intraoperative detection of perforations) were compared for single versus double gloves (including both standard and indicator double gloves), single versus standard double gloves, and single versus indicator double gloves separately.

Outcome 1: Single versus inner double glove perforations

Standard and indicator double gloves combined

Fourteen trials [4, 5, 7, 22,23,24,25,26, 28, 29, 31,32,33,34] with a total of 4781 single and 4658 double gloves were included for comparing this outcome. Studies included had significant heterogeneity (I² = 55%), so a random effect model was used. Across all the included trials, glove perforation rates were lower for inner double gloves (0.0–7.6%) compared to single gloves (0.4–17.6%) with the difference being significant in 12 trials [4, 5, 7, 22,23,24, 26, 28, 29, 32,33,34]. Meta-analysis of data also showed a significantly lower rate of the inner double glove (1.9%) perforations than single glove (8.9%) perforation (OR = 0.2, 95% CI 0.14–0.31, Z = 7.57, P < 0.00001, I² = 55%, P = 0.0006).

Standard double gloves

Analysis of 11 trials [4, 5, 7, 22, 23, 26, 28, 31,32,33,34] that compared single gloves (4021) with standard double (3872) gloves only showed similar results with the inner glove of a standard double glove (2.3%) having significantly lower rates of perforations than single (9.8%) gloves (OR = 0.22, 95% CI 0.15–0.33, Z = 7.27, P < 0.00001, I² = 58%, P = 0.008).

Indicator double gloves

Analysing data from 3 trials [24, 25, 29] with a total of 760 single and 787 indicator double gloves showed a significantly lower perforation in the inner glove (0%) of indicator double gloves than single (4.2%) gloves (OR = 0.06, 95% CI 0.01–0.31, Z = 3.31, P = 0.0009, I² = 0%, P = 0.38).

A subgroup analysis did not show a difference between standard and indicator double gloves (I² = 56.3%, P = 0.13) (Fig. 4).

Fig. 4

Comparison of single glove perforations with inner double glove perforations and a subgroup analysis between standard and indicator double gloves

Outcome 2: Single- versus outer double glove perforations

Standard and indicator double gloves combined

Fifteen trials [4, 5, 7, 21,22,23,24,25,26, 28, 29, 31,32,33,34] with 5493 single and 5205 double gloves were included in this analysis. The rate of the outer double glove (1.2–27.5%) and single glove (0.4–17.6%) perforations was comparable across different trials, and in only three trials difference was significant [28, 29, 34]. The overall rate of outer double glove perforations (10.4%) was also similar to single glove (8.7%) perforations with an OR of 1.09 (95% Cl 0.87–1.37, Z = 0.78, P = 0.4) using random effect model (I² = 57%, P = 0.003).

Standard double gloves

There were no significant differences observed in the rate of the outer double glove (11.7%) and single glove (9.8%) perforations when results were analysed separately including 11 trials [4, 5, 7, 22, 23, 26, 28, 31,32,33,34] which used only standard double gloves (OR = 1.11, 95% Cl 0.86–1.43, Z = 0.81, P = 0.8, I² = 60%, P = 0.005).

Indicator double gloves

A separate analysis of 4 trials [21, 24, 25, 29] using indicator double gloves (1472 single vs. 1306 double) also did not show any significant difference between outer double glove (6.3%) and single glove (5.8%) perforations (OR = 0.96, 95% Cl 0.49–1.91, Z = 0.11, P = 0.92, I² = 59%, P = 0.06).

Subgroup analysis also did not show any difference between the standard and indicator double gloves (I² = 0%, P = 0.7) (Fig. 5).

Fig. 5

Comparison of single glove perforations with outer double glove perforations and a subgroup analysis between standard and indicator double gloves

Outcome 3: Single versus matched inner-outer double glove perforations

Standard and indicator double gloves combined

Thirteen trials [5, 6, 21,22,23,24, 27, 28, 30, 31, 33] with a total of 5930 single and 5396 double gloves reporting this outcome were included in the analysis. Across all the included trials, the rate of matched inner-outer double glove perforations (0–3%) was significantly lower than single glove perforations (5.3–15.2%). The overall rate of matched perforations in double gloves (0.9%) was also significantly lower than single glove perforation (8.4%) with a pooled OR of 0.1 (95% Cl 0.07–0.13, Z = 14.91, P < 0.00001) using fixed effect model (I² = 40%, P = 0.06).

Standard double gloves

A separate analysis of 9 trials [5, 6, 22, 23, 27, 28, 30, 31, 33] using standard double gloves (3816 single vs. 3998 double) including 2 trials [27, 30] that used both standard and indicator double gloves showed similar results with matched perforations in the double gloves (0.07%) reported significantly lower than single glove (8.8%) perforations (OR = 0.07, 95% Cl 0.05–0.11, Z = 13.1, P < 0.00001, I² = 0%, P = 0.57).

Indicator double gloves

Similar results were observed for indicator double gloves where analysis of 4 trials [21, 24, 27, 30] with a total of 2114 single and 1398 double gloves showed a significantly lower rate of matched double glove (1.3%) perforations than single glove (7.6%) perforation (OR = 0.17, 95% Cl 0.1–0.28, Z = 7.01, P < 0.00001, I² = 46%, P = 0.14).

A subgroup analysis showed significantly higher efficacy of standard double gloves in reducing matched perforations than the indicator double glove (I² = 85.7%, P = 0.008) (Fig. 6).

Fig. 6

Comparison of single glove perforations with matched inner-outer double glove perforations and a subgroup analysis between standard and indicator double gloves

Outcome 4: Incidences of hand contamination in single versus double glove condition

Standard and indicator double gloves combined

Data from 6 trials [7, 21, 24, 25, 31, 32] with a total of 1692 single and 1655 double gloves pooled for meta-analysing this outcome. In all the trials, hand contamination rates were lower in the double glove condition (0–3.8%) than single glove condition (0–6.1%), but the difference was significant in only two trials [7, 21]. The overall incidences of hand contamination were significantly lower in double-gloved (0.7%) compared to single-gloved (2.4%) conditions with a pooled OR of 0.28 (95% Cl 0.14–0.54, Z = 3.8, P = 0.0001) using fixed effect model (I² = 0%, P = 0.62).

Standard double gloves

A separate analysis of 3 trials [7, 31, 32] that used standard double gloves (524 single vs. 542 double) showed similar results with significantly lower incidences of hand contamination in standard double-gloved (1.7%) condition than single-gloved (4.8%) condition (OR = 0.34, 95% Cl 0.16–0.75, Z = 2.71, P = 0.007, I² = 0%, P = 0.45).

Indicator double gloves

Meta-analysis of 3 trials [21, 24, 25] that used indicator double gloves (1168 single vs. 1113 double) showed similar results that indicator double gloves (1.7%) are highly effective in protection against hand contamination compared to single gloves (4.8%) with a pooled OR of 0.17 (95% Cl 0.04–0.64, Z = 2.61, P = 0.009) using fixed effect model (I² = 0%, P = 0.73).

A subgroup analysis showed that standard and indicator double gloves are equally effective (I² = 0%, P = 0.36) (Fig. 7).

Fig. 7

Comparison of hand contamination incidences in single glove condition with double glove condition and a subgroup analysis between standard and indicator double gloves

Outcome 5: Intraoperative detection of perforations in single versus double glove condition

Standard and indicator double gloves combined

Eleven trials [6, 7, 21, 24,25,26,27, 29, 30] with a total of 424 perforated single and 456 perforated double gloves were included in this analysis. The glove perforations detected intraoperatively ranged between 5.3 to 100% for single gloves and 9.5 to 100% for double gloves across different trials. Only 4 trials [8, 11, 31, 34] showed a significantly higher percentage of glove perforation detected intraoperatively when double-gloved than single-gloved. The overall intraoperative glove perforation detection was significantly higher in double gloves (58.1%) than in single gloves (33.7%) with a pooled OR of 3.58 (95% Cl 1.43–8.92, Z = 2.73, P = 0.006) using random effect model (I² = 84%, P < 0.00001).

Standard double gloves

Meta-analysis of 5 trials [6, 7, 26, 27, 30] using standard double (250 perforated single vs. 244 perforated double) including 2 trials [27, 30] which used both standard and indicator double gloves showed that standard double gloves have similar efficiency of intraoperative glove perforation detection (32.8 to 38.9%) as single gloves (OR = 1.61, 95% Cl 0.53–4.89, Z = 0.83, P = 0.4, I² = 80%, P = 0.0004).

Indicator double gloves

Analysis of 6 trials [21, 24, 25, 27, 29, 30] using indicator double gloves (174 perforated single vs. 212 perforated double) separately showed that double-indicator gloves (80.2%) had significantly higher efficiency of intraoperative perforation detection than single gloves (35.1%) (OR = 8.64, 95% Cl 4.78–15.61, Z = 7.15, P < 0.00001, I² = 14%, P = 0.33).

A subgroup analysis showed significantly higher efficacy of indicator double gloves than standard double gloves (I² = 85.4%, P = 0.009) (Fig. 8).

Fig. 8

Comparison of intraoperative detection of perforations in single glove condition with double glove condition and a subgroup analysis between standard and indicator double gloves

Sensitivity analysis

For the outcome of single versus double inner double glove perforation, a reanalysis of six trials [5, 7, 25, 28, 31, 32] having either low or unclear risk of bias showed a similar significant OR and heterogeneity (OR = 0.17, 95% Cl 0.08–0.37, I² = 63%, P = 0.02) as in the overall result (OR = 0.2, 95% CI 0.14–0.31, I² = 55%, P = 0.006). For the outcome of single versus matched double glove perforation also, a reanalysis of seven trials [5, 6, 21, 27, 28, 31] did not show any major change in OR and heterogeneity (OR = 0.1, 95% Cl 0.07–0.15, I² = 34%, P = 0.16) than the overall result (OR = 0.1, 95% CI 0.07–0.13, I² = 40%, P = 0.06). For the outcome, incidences of hand contamination in single versus double gloving, five trials were reanalysed [7, 21, 25, 31, 32] and showed similar OR and heterogeneity (OR = 0.28, 95% CI 0.14–0.54, I² = 0%, P = 0.45) as in the overall result (OR = 0.28, 95% Cl 0.14–0.54, I² = 0%, P = 0.62). The same is observed for the outcome, intraoperative glove perforations, where reanalyzing five trials [6, 7, 21, 25, 27] did not affect OR and heterogeneity much (OR = 4.75, 95% Cl 1.22–18.53, I² = 79%, P = 0.0006) and is similar to overall result (OR = 3.58, 95% Cl 1.43–8.92, I² = 84%, P < 0.00001).

For three outcomes (inner double glove perforations, outer double glove perforations, and intraoperative detection of perforations), heterogeneity among studies was found to be more than 50%. For the 1st and 2nd outcomes excluding Makama et al. (2016) [28], the heterogeneity (I²) dropped from 55 to 37% and from 57 to 23%, respectively, and OR remained similar (1st outcome: OR = 0.24, 95% Cl 0.17–0.35; 2nd outcome: OR = 1.05, 95% Cl 0.86–1.24). For 3rd outcome, excluding any of the articles did not have much effect on heterogeneity.

Subgroup analysis

For an outcome that has significant results and a maximum number of trials included (single vs. inner double glove perforations), we did three subgroup analyses. We analysed the outcome for the economic status of countries and found that double gloving was slightly more effective in HICs (OR = 0.16, 95% Cl 0.08–0.34) [4, 24,25,26, 29, 32, 34] than LMICs (OR = 0.22, 95% Cl 0.13–0.39) [5, 7, 22, 23, 28, 31, 33], but the difference was nonsignificant (I² = 0%, P = 0.51) (Fig. 9). In trials conducted after 2000 [4, 5, 7, 24, 28, 33] double gloving was slightly more effective (OR = 0.13, 95% Cl 0.07–0.26) than in those conducted before 2000 (0.27, 95% Cl 0.17–0.42) [22, 23, 25, 26, 29, 31, 32, 34]. The subgroup difference was nonsignificant (I² = 66.2%, P = 0.09) (Fig. 10). We observed that those having surgery duration < 2 h [4, 23, 25] double gloving provided slightly higher protection (OR = 0.13, 95% Cl 0.05–0.37) than in trials where surgical duration was > 2 h (OR = 0.34, 95% Cl 0.19–0.64) [5, 31], but the difference was nonsignificant (I² = 59.2%, P = 0.12) (Fig. 11).

Fig. 9

A subgroup analysis between trials conducted in low- and middle-income countries and high-income countries for the outcome single glove perforations versus inner double glove perforations

Fig. 10

A subgroup analysis between trials conducted before year 2000 and after 2000 for the outcome single glove perforations versus inner double glove perforations

Fig. 11

A subgroup analysis between trials having an average surgical duration of less than 2 h and more than 2 h for the outcome single glove perforations versus inner double glove perforations

Publication bias

There seems to be no publication bias for any of the outcomes as the funnel plot was close to symmetric in each case (Fig. 12). The result of Egger’s test was also nonsignificant.

Fig. 12

Funnel plot for assessing publication bias in five outcomes: A single versus inner double glove perforations, B single versus outer double glove perforations, C single versus matched inner-outer double glove perforations, D incidences of hand contamination in single versus double glove condition, and E intraoperative detection of perforations in single versus double glove condition

Discussion

Summary of outcome measures and explanations

In this systematic review and meta-analysis, we observed that double gloving, either standard or indicator, gives significant protection over single gloving in terms of reduced inner glove and matched perforations and lower incidences of hand contamination. Indicator double gloves are effective for intraoperative detection of perforations but not standard double gloves. It cannot be said that detecting perforations early due to indicator gloving makes surgeons change their behaviour in favour of getting fewer perforations as matched glove perforations are significantly lower in standard than indicator glove group, and other outcomes are similar for both standard and indicator double glove groups. Further, the better efficacy of standard double gloves in reducing matched glove perforations than indicator double gloves observed in this study may indicate standard double gloves being more protective than indicator double gloves. However, it is not conclusive as other outcomes of this study do not validate this interpretation and show equal efficacy of standard and indicator gloves in reducing perforations. Moreover, no previous study is available to support this conclusion. Indicator gloves are better at intraoperative detection of perforations which may motivate the surgical personnel to change their gloves, thereby reducing their risk of percutaneous injuries and hand contamination [35]. Given that evidence on whether standard or indicator double gloves provide better protection to surgical personnel remains inconclusive, their cost-effectiveness is an important consideration to adopt one over the other. Though there is no study evaluating the cost-effectiveness of the indicator over standard double gloving in surgeries, according to price listings at different marketplaces, indicator gloves are much costlier than standard double gloves owing to their technological aspect. Standard double gloves might be a more feasible option for already resource-constrained LMICs. However, future research is needed to weigh the cost and benefits of standard and indicator double gloving in surgeries. This would provide practical insights for policymakers and healthcare providers who need to balance safety with budget constraints.

This systematic review also showed that deep/major surgeries, emergent surgeries, principal surgeons, and longer duration of surgery are major risk factors for high glove perforation rates. We observed that double gloving provides slightly better protection where the surgical duration was < 2 h than where it was > 2, but the difference was nonsignificant which could be due to the small sample size (only five trials). These results are supported by the fact that more extensive surgical exposure is associated with a higher risk of glove perforations and contamination and needs better protective gloving which double gloving may provide [1]. We also observed that impaired dexterity is not a major constraint against double gloving and does not impact glove perforation rates as outer glove perforations are similar in both single- and double-gloved groups.

Double gloving was found to be slightly more effective against single gloving in HICs and trials conducted after 2000 compared to LMICs and trials conducted before 2000. Better surgical environments and better compliance to double gloving in HICs and recent trials could be the reason for it. It is also reflected in WHO data where surgical personnel from LMICs have a higher rate (10–65%) of occupational exposure to infections (HBV/HCV/HIV) than HICs (0.5–27%) [2].

For a few outcomes (inner/outer double glove perforations, intraoperative detection of perforations), the heterogeneity among trials was very high (> 50%) indicating methodological differences as they use only one or other method of perforation detection, and information on the methods of randomization, allocation concealment, and blinding was not provided for a majority of trials. The heterogeneity for two outcomes (inner and outer glove perforations) was reduced significantly by excluding Makama et al. (2016) [28]. The reason could be the poor methodological quality of this trial as no information was provided for a method of randomization, blinding of outcome assessor, and which gloves are included or excluded. The overall high risk of bias in trials does not seem to influence outcomes as excluding these trials does not cause any significant changes in the outcome. However, in a majority of trials, three risks-of-bias domains chosen for overall risk of bias categorizations are in an unclear category, and it cannot be conclusively said whether the overall risk of bias has an actual impact on outcomes or not.

Comparison with other studies

The results are in line with many other reviews and meta-analyses despite the addition of many new trials in this analysis. Where a significant reduction in inner glove perforations and no effect on outer glove perforation was reported [9,10,11] for both indicator and standard gloves [16, 36] when compared with single gloves. In previous reviews, also indicator gloves were found to be more effective for intraoperative detection of perforations than standard gloves [9, 10, 36]. Studies also have reported significantly reduced matched glove perforations when double-gloved [11, 16], but only one meta-analysis compared standard and indicator gloves for this outcome with only one trial and found no difference [16]. Few reviews also reported a significant reduction in blood contamination incidents when double-gloved [10, 16], but they included only two or three trials in the analysis. None of these compared the outcome for standard and indicator gloves separately. One meta-analysis compared inner glove perforation outcomes for LMIC and HICs [16], and another compared it for surgical duration < 2 h and > 2 h [11] and reported it to be slightly lower in HICs and when the surgical duration was < 2 h respectively as we found in our comparison.

Limitations in the quality of evidence

Most of the trials included in this meta-analysis have poor methodological quality. The majority of trials used only one method of perforation detection, and neither of these was reported to have 100% efficiency and accuracy [32]. Even those trials that used standard water leak tests recognized by the American Society for Testing and Materials and the European Standards Committee also did not use the recommended 1000-ml volume of water. Because the methodology was not clearly described for many trials while extracting the data, several calculations and logical conclusions have to be made based on the details provided in the study. Like in many trials, it was not clear if perforations were reported in terms of the number of perforated gloves or the number of perforations [5, 23, 31, 32]. From the details provided in the study, number of gloves seemed more appropriate unit and taken as such. Moreover, it applies equally to all the groups in the study, and it will not impact the outcome. In a few trials, it is not clear if gloves are given as a pair or set or total number. In some trials, there is no clear information on the number of gloves in each group [31, 32]. Like in the study by Gani et al. (1990), the total number of gloves in each group was calculated from the number and percentage of perforated gloves provided [26]. Thomas et al. (2001) gave only the total number of gloves and did not specify anything, so we divided the total by 3 to assign gloves equally in single, outer, and inner gloves [31]. Doyle et al. (1992) also does not mention the number of single or double gloves; instead, the number of bags with valid questionnaires in which gloves are put was taken as the total pairs of single or double gloves [32]. In three trials, instead of matched perforations, both inner and outer glove perforation terms were used [6, 27, 30], which implies that perforations are in pairs, but it is not explicitly clear if the perforations are at the same location on inner and outer gloves. In 1 trial, only 1 incidence of needlestick injury was counted as exposure as all other 18 that may lead to exposure in a single gloving group as per the study were not visually or experimentally authenticated [24]. For perforated double gloves detected intraoperatively, few trials only reported perforated double gloves and did not specify anything whether total, inner, or outer [6, 7, 27, 30], while few trials did not specify whether these perforations were out-of-total double glove perforations (inner and outer included) or out of inner or outer glove only, and we considered it as out-of-total double glove perforations [6, 27, 30]. Further details were provided in the supplementary data table.

In this study, many of the confounding factors also could not be taken care of as most of the trials included were performed in a mix of different surgical specialties, involved a combination of surgical personnel, and had a great variation in surgical duration. Including only English-language studies is another limitation which may potentially affect the generalizability of the results. In the majority of trials, information on randomization, allocation concealment, and blinding could not be retrieved, which may be a limitation for the strength of evidence.

Conclusion

Based on the results, we recommend the use of double gloves over single gloves for protection of surgical personnel from perforations and contaminations. Indicator glove use was recommended for better detection of perforations intraoperatively and to make surgeons aware to change them timely; however, it does not provide additional protection over standard double gloves, and there should be an evaluation of its cost-effectiveness over standard double gloves.

Also, the different surgical specialties and different surgical personnel as well as the duration of surgery have different risks of perforations. Different high-quality trials should be conducted in the future especially to compare their effectiveness for different surgical specialties, type of surgical personnel, and duration of surgery, so that specific recommendations can be made.