Effects and safety of atmospheric low‐temperature plasma on bacterial reduction in chronic wounds and wound size reduction: A systematic review and meta‐analysis

1. INTRODUCTION

Highly energetic physical plasmas comprise a mixture of reactive ionised non‐thermal particles containing diverse biologically reactive factors including charged particles, free radicals, excited atoms and molecules, photons, and electromagnetic fields, which present as “cold or low temperature plasma” at atmospheric pressure.1 Low‐temperature atmospheric plasmas (APs) are generated under atmospheric pressure at ambient temperatures ranging from 20°C to 50°C. With the development of low‐temperature APs, at temperature ranges of approximately 38°C at the point of application,2 new therapeutic options directed against prokaryotic cells (eg, microorganisms) living on eukaryotic cells (eg, human tissue) are available.3

Interest in the medical application of APs is rapidly increasing. The first study on the use of argon plasma for tumour removal were reported in 1989.4 Their potential therapeutic benefits were later explored for the treatment of chronic wounds,5, 6, 7, 8 ablation of non‐neoplastic Barrett mucosa,9 and other neoplastic disease.10, 11, 12, 13, 14, 15

Although the majority of published studies reported results obtained from laboratory‐based experimental work,16, 17, 18, 19, 20 a number of clinical trials involving patients treated with AP for neoplastic9, 21 or skin disease22 have been reported. The management of acute and chronic wounds has emerged as one of the promising indications for the clinical use of APs because of their experimentally demonstrated properties, which have been shown to improve healing of stagnating, chronic, open wounds and to reduce bacterial burden in colonised or infected wounds.23 Two distinct features support the use of APs to treat or prevent infection, namely, their demonstrated in‐vitro antimicrobial effectiveness—even against bacterial spores (depending on application time and physical parameters)—and their remarkable access into narrow and confined spaces and structures.24, 25, 26, 27, 28 In light of the continued development of bacterial resistance against antibiotics, non‐antibiotic‐based methods to manage colonised or infected wounds, and simultaneously to promote wound healing, theoretically appear to be even more attractive. With such technology producing direct or indirect low‐temperature APs on viable tissue, it could be possible to directly decontaminate patients' wounds, which are colonised with pathogenic or potentially pathogenic microorganisms.29 Finally, APs could also have the potential to be used to deliver drugs, including antimicrobial active compounds, into deeper layers of tissue or difficult‐to‐reach anatomical regions.

The aim of this systematic meta‐analysis was to screen existing randomised trials that have studied the use of APs to promote chronic wound healing, reduce the bacterial burden in wounds, and to determine the safety of AP application.

2. METHODS

The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analysis) guidelines were followed.30

3. RESULTS

An initial search identified 96 studies. Thirteen studies were assessed as being suitable for full review. Nine studies were eligible for full critical appraisal and were therefore included for further analysis. Among the nine identified studies, eight were RCTs, and one was a prospective cohort study. The detailed process of selection is summarised in Figure 1. Nine studies encompassing 268 patients, randomised either to treatment with AP or control, were identified.6, 8, 36, 37, 38, 39, 40, 41, 42 Four studies investigated the effect of AP on wound healing compared with conventional standard treatment,8, 36, 38, 40 and five studies presented data on reduction of bacterial burden in wounds after AP application.6, 8, 32, 36, 39 All nine studies reported the occurrence of SAEs in the intervention and control arms of the respective studies.

There was substantial heterogeneity in the study protocols. Primary differences were the patient selection characteristics, the plasma source used, and the technical and physical specifications of the plasma application. Study characteristics are summarised in Tables 1, 2, 3.

The results of the risk‐of‐bias evaluation are presented in Table 4. Overall, there was serious risk of bias, predominantly because of unclear or high risk of selection and performance bias. There was an insufficient number of studies included in the separate meta‐analyses for appropriate interpretation of the funnel plots. The bias of using different physical parameters of the plasma sources was impossible to estimate as they were inconsistently stated in the analysed studies, which is partly explained by the included studies having used different AP devices. Two studies37, 38 used the MicroPlaSter β plasma torch (ADTEC Plasma Technology Co. Ltd., Hiroshima, Japan), two further studies39, 40 used MicroPlaSter α plasma torch (ADTEC Plasma Technology Co. Ltd., Hiroshima, Japan), one trial6 used both the MicroPlaSter α or β plasma torch (Table 2), two RCTs8, 42 used kINPen Med (Neoplas tools GmbH, Greifswald, Germany), and one trial36 used PlasmaDerm VU‐2010 (Cinogy GmbH, Duderstadt, Germany). AP application times ranged from 60 to 300 seconds (mean ± SD = 120 ± 98 seconds), and all but one study36 used Argon gas flow at various gas flows ranging from 2.2 slm6, 37, 38, 39, 40 to 5.0 slm.8, 42 The AP power density was stated in only one study,36 being 120 mW/cm².

Comparisons, corresponding data, and meta‐analyses are presented as forest plots in Figures 2, 3, 4. Four RCTs found the wounds of patients treated with AP not to show a significant improvement in size (AP vs control: Figure 2) compared with wound treatment that did not involve AP.8, 36, 38, 40 Five RCTs6, 8, 36, 39, 42 were identified that investigated a bacterial reduction using different assessment criteria among the individual studies. Patients treated with AP did not show an improved reduction in bioburden compared with controls (AP vs control: OR = 0.85; 95% CI = 0.45‐1.62; I ² = 67%; P = 0.63, Figure 3).

All nine RCTs reported on the occurrence of SAEs. Among the pooled 268 patients, only 1 patient treated with AP developed an SAE,36 as well as 1 control patient36 who was not treated with AP (OR = 1.00; 95% CI = 0.05‐19.96; P = 1.00, Figure 4). The patient who developed a SAE in the plasma‐treated group was hospitalised because of backache as a result of a vertebra shift on basis of pre‐existing osteoporosis. This SAE was labelled as being unrelated to plasma treatment.

GRADE tables with full assessment of the individual comparisons are presented in Table 5. Overall, the quality of evidence was assessed as being low to very low related to risk of bias and imprecisions of analysed studies.

4. DISCUSSION

This is the first systematic review and meta‐analysis to explore the clinical effectiveness of AP used in patients with open, chronic skin wounds. GRADE methodology was used to assess the quality of the retrieved evidence. Overall, the quality of the evidence included in this systematic review and meta‐analysis is moderate to very low because of the serious risk of bias and serious imprecision of identified studies. Based on the analysed data, evidence shows that the application of AP in wound care is safe, yet it is ineffective in the reduction of wound size or bacterial bioburden on wounds compared with other treatment modalities.

However, it is important to note that the available studies comparing AP with treatment modalities without application of AP used different and difficult‐to‐compare AP sources and application modalities. The chemical composition and the physical characteristics of the generated AP depend on a number of variables such as pressure, gas mixture, design of the device, physical stimuli, and surrounding environmental factors. Therefore, different AP sources are difficult to compare with each other, and the results of this meta‐analysis are interpreted with great caution. Future studies should include technical details of the applied AP sources, including—at a minimum—information on gas mixture and gas flow rate; voltage; and, if applicable, amplitudes of alternating voltage pulses, power density, UV spectrum, direct or indirect AP source built type, distance to surface, and application time.

The present meta‐analysis includes one cohort study41 and eight prospective randomised trials. While RCTs are usually the focus of a meta‐analysis because of the least risk of bias, the same methodology used for randomised trials can be applied to cohort studies.43 Therefore, and because of the data structure supporting the research questions of this meta‐analysis, the authors decided to include this cohort study into the analysis.

We focused on three main study outcomes: the occurrence of SAEs, the ability of AP to reduce the size of chronic wounds and to reduce microbial burden in wounds. While all studies used the same definitions for SAEs, the definitions for reduction of wound size or reduction of the bacterial load in wounds were different. Four studies8, 36, 38, 40 contained sufficient information to analyse the effect of AP on wound healing compared with conventional standard treatment, and five studies provided data on the reduction of bacterial burden in wounds after AP application compared with controls.6, 8, 36, 39, 42

None of the investigated studies provided methodically identical and comparable results for both outcome measures. Wound size measurements were undertaken with either a ruler, a transparent film with printed squares to draw wound borders and count the number of squares, or a technical measurement device (Visitrak; Smith and Nephew Healthcare, Hull, UK). Furthermore, there were inconsistencies on the time intervals for measuring wound size, but most studies included information on wound size at the start of the study and after 14 days of treatment. As all four studies that measured wound size presented results differently, a decision was made to compare the efficacy of AP or control treatment to reduce the wound size based on the outcome allocation of the individual studies. For instance, Brehmer et al36 reported “a more than 50% reduction in ulcer size … in 5/7 and 4/7 patients in the standard and plasma group, respectively.” However, if absolute values, for example, reduction of wound size in cm² after a defined treatment period, would have been used to assess efficacy, a more pronounced ulcer size reduction would have been observed in the AP group compared with the standard group until the end of the treatment period at visit 21 (standard group: −3.4 cm² vs AP group: −5.3 cm²). In the same year, Ulrich et al8 reported a 12.5% reduction of wound size from 14.1 ± 12.2 (mean ± standard deviation) cm² to 11.6 ± 10.2 cm² in the control group and a 39% reduction in the AP group (from 4.4 ± 4.3 cm² to 2.9 ± 3.3 cm²) over a 14‐day study period. Such differences in the study methodologies may also explain the heterogeneity (I ² = 12%) of the pooled outcome results for wound size reduction.

Similarly, antibacterial outcome measures were reported with different scales, and we could not pool the related data to obtain a more powerful conclusion. Therefore, and because setting clinically relevant thresholds for bacterial reduction in the context of wound care are debatable, we have again used the original assessment criteria for efficacy as defined and used by the individual studies. Using bacterial reduction as an outcome variable to assess the clinical relevance of an antimicrobial method is controversial.44 For instance, Isbary et al39 reported a “significant reduction in bacterial counts” in the AP‐treated group. Indeed, after application of AP, a mean 1.10 log₁₀ reduction was observed in the intervention group and a 0.41 log₁₀ reduction in the control group. Although the difference in the mean reduction between the two study arms is significant, an intervention achieving a 1 log₁₀ bacterial reduction would hardly be regarded as relevantly “bactericidal” or “antimicrobial.”

Indeed, log₁₀ reduction data for antimicrobial compounds or interventions on wounds are difficult to compare because of the widely differing methodologies and antimicrobial concentrations/application times used in various methods. Although there is no internationally accepted definition for an antimicrobial device, there is a general agreement to identify a compound or a device as “bactericidal” if it is demonstrated to produce at least a 3‐log₁₀ reduction in the number of tested viable bacterial cells.45, 46, 47 However, dilution tests or time kill‐kinetics do not necessarily represent conditions found in wounds. In‐vitro models represent non‐competing environments, and significant log₁₀ reductions can be achieved with low concentrations of antimicrobials against mostly planktonic test strains. No account is made of phenotypically different persister cells present in biofilm present in all chronic wounds, which require sophisticated methods for identification.47 Therefore, demanding a 3‐ or 4‐log₁₀ reduction of microbial cells in a wound may represent unrealistically high expectations. Yet, a 1‐log₁₀ reduction would be so minute that it would not be possible to distinguish a decontaminating (eg, removing bacteria) from a disinfecting (killing or inactivating bacteria) effect. Therefore, we did not apply the term “bactericidal” or “antibacterial” but used the expression “bacterial reduction.”

Finally, it must be pointed out that, among the identified trials comparing the bacterial reduction of AP against treatment modalities without AP one trial, Ulrich and colleagues8 used a control study arm in which a combination of 0.1% octenidine‐dihydrochloride and 2% 2‐phenoxyethanol (OCT; Octenisept, Schuelke GmbH, Norderstedt, Germany) was applied.48 The statistical design of this RCT was a non‐inferiority study showing that AP achieved a significantly inferior bacterial reduction compared with the wound antiseptic used as a control. The weight of this study in the context of the conducted meta‐analysis was 15.7% and could have potentially biased the outcome results for bacterial reduction in favour of the control. However, a subset analysis of all RCTs providing data for bacterial reduction, excluding the study with octenidine‐dihydrochloride as control, showed that there was no difference in the results irrespective of whether AP was compared against comparable controls or all controls, including the RCT with an antiseptic as control (AP vs control excluding Ulrich et al8: OR = 0.69; 95% CI = 0.21‐1.08; I ² = 84%; P = 0.10).

To our knowledge, there is only one more RCT that compared AP against OCT or the sequential application of AP followed by OCT treatment on chronic, open wounds.49 This three‐arm RCT reported that the application of AP or OCT result in similar “microbial reduction classes,” with no further reduction if both treatments were applied sequentially on wounds. At first glance, the finding of Klebes et al49 appears to be in direct contrast to the results reported by Ulrich et al.8 However, both RCTs used absolutely not comparable microbiological measurements. Because of this and other limitations, we therefore could not include this latter RCT into our meta‐analysis as, in most of the 34 included patients, the three treatment procedures were performed on each wound, and only the immediate antibacterial effect of the applied interventions were investigated and reported semi‐quantitatively as “4—abundant,” “3—moderate growth,” “2—little growth,” “1—marginal growth,” and “0—no growth.” Finally, changes of wound size and occurrence of SAEs were not included as outcome measures in this study.49

In conclusion, patients treated with or without AP had similar outcomes of wound size reduction and bacterial reduction. Use of AP for treatment of wounds is safe, but the current evidence shows that there is no clinical benefit of AP in wound care using currently applied physical plasma parameters. Future studies must include technical details of the applied AP sources and should use comparable outcome measures based on reproducible definitions.

Assadian O, Ousey KJ, Daeschlein G, et al. Effects and safety of atmospheric low‐temperature plasma on bacterial reduction in chronic wounds and wound size reduction: A systematic review and meta‐analysis. Int Wound J. 2019;16:103–111. 10.1111/iwj.12999