Evaluation of three experimental in vitro models for the assessment of the mechanical cleansing efficacy of wound irrigation solutions

Abstract

The aim of this study was to compare different wound‐rinsing solutions to determine differences in the efficiency and to evaluate three different in vitro models for wound cleansing. Different wound‐rinsing solutions (physiological saline solution, ringer lactate solution for wound irrigation, water and a solution containing polihexanide and the surfactant undecylenamidopropyl‐betain) were applied on standardised test models (one‐ and three‐chamber model, flow‐cell method and a biofilm model), each challenged with three different standardised wound test soils. In the one‐chamber model saline showed a better effect on decontaminating proteins than the ringer lactate solution. In the flow‐cell method, water performed better than physiological saline solution, whereas ringer lactate solution demonstrated the lowest cleansing effect. No obvious superiority between the two electrolyte‐containing solutions was detectable in the biofilm model. Unfortunately, it was not possible to assess the protein decontamination qualities of the surfactant‐containing solution because of the interference with the protein measurement. The flow‐cell method was able to detect differences between different rinse solutions because it works at constant flow mechanics, imitating a wound‐rinsing procedure. The three‐chamber and the less‐pronounced modified one‐chamber method as well as the biofilm model had generated inhomogeneous results.

Introduction

Wound cleansing is an imperative component of wound bed preparation and is a standard goal during care of acute and chronic wounds 1.

Wound cleansing involves the use of liquid cleansers and, depending on the method of administration and the duration and vigour of treatment, removal of dirt, crusts, dead tissues, biofilm, microorganisms and debris. If an antiseptic efficacy is indicated in addition to mechanical cleansing, that is, for dirty traumatic soft tissue wounds, rinsing solutions containing an antiseptic agent should be preferred 2, 3. Traditionally, two techniques are used for wound cleansing 4, 5: either rinsing the wound under pressure, for example, with a syringe using sterile irrigation solutions 6, 7, or mechanically wiping the wound with a moist sterile cloth 8. In the past years, new technologies using high‐pressure lavage or negative pressure wound therapy with instillation have enlarged the spectrum of application methods of wound‐cleansing solutions on wounds 9.

However, irrespective of which method is applied, the selection of the cleansing solution remains controversial. Today, a variety of cleansing solutions exist, including non‐antimicrobial cleansing solutions such as water, isotonic saline (0·9% saline solution) or ringer lactate solution and a cleansing solution containing a surfactant, which also contains an antimicrobial preservative in antiseptically effective concentration.

Either way, the selection of wound cleansers should be based on their cleansing efficacy and their lack of cytotoxicity. While cytotoxicity can be measured using international harmonised in vitro standards 10, the assessment of the cleaning effect of various cleansing solutions – specifically in a biofilm environment – requires separation of two distinct elements: a physical–mechanical and a chemical component. So far, no standard method exists for the quantitative assessment of the mechanical and chemical cleaning effects of rinsing solutions 11. A three‐chamber model and a flow‐cell model are described as models to determine the cleansing efficacy of wound irrigation solutions 12, 13. Furthermore, some models allow assessment of specific parameters, such as the hydrodynamic conditions of different rinsing pressures 14, 15. However, these methods have not been used to evaluate the cleansing effect of different wound‐cleansing solutions in the presence of test soil simulating wounds. Therefore, using these methods, the cleaning effect of different wound‐rinsing solutions was determined experimentally and compared in order to identify a suitable test model for further studies.

Methods

All experiments were conducted under identical conditions, only differing in the tested cleansing solution in order to allow comparability. The following cleansing solutions were included: (i) deionised water, (ii) 0·9% isotonic saline solution (0·9 g saline to 100 ml; Karl Roth GmbH, Karlsruhe, Germany), (iii) ringer lactate solution (B. Braun, Melsungen, Germany) and (iv) 0·1% undecylenamidopropyl betain with 0·1% polihexanide (Prontosan^® wound irrigation solution; B. Braun, Sempach, Switzerland) used as a preservative at an antiseptically effective concentration 2.

Preparation of the test soiling (three‐chamber slides)

Sample slides (14 mm slides; Thermo Fischer Scientific, Braunschweig, Germany) for the three‐chamber model were washed before use with 2 ml of alkaline sodium dodecyl sulphate (SDS) solution (pH 11) and depurated in an ultrasonic bath for 10 minutes with 2 ml of double‐distilled water. The preparation of the plasma slide (P‐slide) and the fibrin slide (F‐slide) was performed by adding fresh frozen plasma (anticoagulated with citrate and stabilised with phosphate and dextrose, CPD plasma; Institute of Transfusion Medicine of the University of Greifswald) thawed at room temperature or 0·9 ml CPD plasma coagulated with 0·1 ml 1·8 M calcium chloride solution, respectively. For both test soils, each field of the three‐field slides was coated with 0·07 ml of the soil and was dried overnight under the laminar flow cabinet.

Preparation of the 24‐well micro‐titre plates (biofilm model)

Pseudomonas aeruginosa SG81 (isolated from a technical water system of the University of Duisburg‐Essen, Aquatic Microbiology) 16 was pipetted into a well of the 24‐well micro‐titre plate, covered with a plastic sheet and incubated at 37°C for 48 hours. On days 3–7, a medium exchange was carried out daily, in which a total of 0·5 ml of the culture medium was added. The biofilm micro‐titre plate was incubated on a shaker at 240 rpm after medium change at room temperature. On the day of the cleansing test (day 8), the medium was aspirated with a pipette, and the resulting biofilm was washed 1× with phosphate‐buffered saline (PBS).

Additionally, wound cleansing solutions were tested using either Test Object Surgical Instruments (TOSI^®), a commercially available test soil simulating human blood on metal strips; DIN ISO 15883 (PEREG GmbH, Waldkraiburg, Germany); or TOSI^® Gold, a commercially available test soil with denaturised proteins on metal strips (PEREG GmbH, Waldkraiburg, Germany).

Three‐chamber model

The cleaning procedure on the three‐chamber slides was performed in histological staining troughs. Briefly, 20 ml of the tested solution heated to 36°C was applied, and three three‐chamber slides (P‐slide or F‐slide) were added. The solution including the three‐chamber slides was stirred manually 3 times every 5 minutes for altogether 60 minutes. For every experiment, a negative control (tested solution without slides) was established. Protein measurement was carried out after 2, 30 and 60 minutes by removing 1·5 ml rinsing solution. The quantitative measurement of residual proteins on slides was conducted after 3x rinsing with 1·5 ml of alkaline SDS (1%). Crystal violet (CV) dye was used for the visual quantification of residual proteins. Additionally, the three‐chamber model was modified as a one‐chamber model using just one field contamination area. Here, the cleaning procedure was carried out in a 25‐ml beaker glass (instead in the staining troughs) following the same procedure as described for the three‐chamber model.

Biofilm model

The biofilms in the 24‐well micro‐titre plates were rinsed once with 1 ml of 0·9% saline solution before the cleansing experiments in order to remove the culture medium and to expose the biofilm. Briefly, 2 ml of the test solutions (ringer lactate solution, isotonic saline solution or the surface active cleaning solution) was added to the wells. The performance of the rinse solutions on the biofilm was determined after 2 minutes and after 30 minutes. In each run, all three wound irrigation solutions were tested in parallel on a 24‐well micro‐titre plate.

Flow‐cell method

A flow cell (Borer Chemie AG, Zuchwil, Switzerland) operated with a pump (microperpex peristaltic pump, LKB, BROMMA, Sweden) was used to apply the respective test solution at a flow rate of 1 ml/minute through the flow cell equipped with the test specimen (Figure 1). The effluent solution was collected during the first 4 minutes in 0·5 ml fractions and then every minute in 1 ml fractions in Eppendorf centrifuge tubes. Overall, one test required 15 minutes.

Figure 1.

Flow diagram of the flow‐cell model with inlet and outlet hoses (internal diameter 3 mm) showing the test chamber with TOSI^® test soil.

Protein detection methods

Bicinchoninic acid (BCA) assay: 1 ml of rinsing fluid was mixed with 2 ml BCA reagent and incubated for 30 minutes at 60°C. After cooling in a water bath, extinction was measured at 562 nm (calibration standard bovine serum albumin 25, 50, 75, 100, 125 and 150 μg/ml) 17.

Roti‐Quant universal kit: 1·5 ml of rinsing fluid was centrifuged at 2·500 U/minute; the protein concentration was determined in 0·1 ml of the supernatant using the Roti‐Quant universal kit (Karl Roth GmbH, Karlsruhe, Germany) according to the manufacturer's instruction 18.

Statistical analysis

Statistical data analysis was conducted by variance analysis (ANOVA), two‐sided and post hoc test using the Tukey test. The software GraphPad Prism Version 6 (GraphPad Software, Inc., San Diego, California, USA) was used to draw the graphs of the individual measurements.

Results

Three‐chamber and one‐chamber models

The total protein quantities applied to the slides ranged from 1500 μg protein/ml to 1872 μg protein/ml (F‐slides) and 1750 μg protein/ml to 1876 μg protein/ml (P‐slides). No reproducible results were obtained using the three‐chamber model, yet the modified one‐chamber model performed better (Figures 2 and 3). However, both methods showed no statistically significant differences between the rinsing solutions on the P‐slide. On the F‐slide, however, the solution containing a surfactant was significantly more effective (P < 0·0001) after 2, 30 and 60 minutes rinsing time. For both slides, a statistically significant difference was observed for the increase of protein concentration in the effluent rinsing solution over time (for ringer lactate solution and a solution containing a surfactant, P < 0·05; for NaCl, P < 0·01).

Figure 2.

Purge capacity, determined as protein concentration after rinse with saline solution (blue), ringer lactate solution (red) and solution containing surfactant (green) on plasma‐contaminated slides (each n = 6).

Figure 3.

Purge capacity, determined as protein concentration after rinse with saline solution (blue), ringer lactate solution (red) and solution containing surfactant (green) on fibrin‐contaminated slides (each n = 9).

In addition to the protein concentration, the total protein amount was determined as the sum of protein in the effluent rinsing solution and the residual protein quantity remaining on the slides after rinsing them. The total protein amount varied considerably between different test solutions (Figure 4). Surprisingly, the surfactant‐containing cleansing solution in the BCA method as well as in the semi‐quantitative determination using CV strain assay demonstrated the highest quantity of residual proteins on surfaces. Both the P‐ and the F‐slides showed residual proteins after rinsing with saline solution and the surfactant‐containing cleansing solution (Figure 5).The proteins were completely removed only after a third rinse with hydrochloric acid/ethanol, that is, the CV staining was negative on all slides.

Figure 4.

Proteins in the rinse solution, residual protein quantity on fibrin‐contaminated slides and total protein quantity after the cleaning procedure with different cleansing solutions.

Figure 5.

Detection of residual protein with crystal violet (CV) staining ‐ Visualisation of the residual proteins by CV staining of the P‐slides (left) and F‐slides (right) after the purification procedure with saline solution (first row) and surfactant solution (second row).

Biofilm model

In the first experiment, after a 2‐minute rinse, the protein concentration in the effluent rinsing solution was 497 μg protein/ mL for the surfactant‐containing cleansing solution, 352 μg protein/ml for the saline solution, 340 μg protein/ml for the ringer lactate solution and 334 μg protein/ml for deionised water. For all rinsing solutions, there was an increase in the protein concentration in the effluent rinsing solution after 30 minutes. However, repeating the experiment under equal conditions, the measurements yielded considerably different results. Therefore, in order to obtain comparative data, the experiments with the biofilm model were carried out in five measurement series on four consecutive days. The repetitions (n = 20) showed a strong variation of the results at the individual test days (Table 1). Taking all measured values into account, there were no significant differences between the protein concentration in the effluent rinsing solution of different rinse solutions and, thus, in their cleaning performance in the biofilm model.

Table 1.

Experiments with different rinsing solutions in the biofilm titre plate model over four consecutive days (n = 20)

Experiment	1 (n = 5)		2 (n = 5)		3 (n = 5)		4 (n = 5)		1–4 (n = 20)
Purging time	M (μg/ml)	SD	M (μg/ml)	SD	M (μg/ml)	SD	M (μg/ml)	SD	M (μg/ml)	SD
0·9% saline solution
2 minutes	352	15	27	4	234	33	133	27	187	125
30 minutes	677	149	32	4	180	90	202	18	273	261
Residual proteins	305	26	10	3	24	29	77	25	89	116
Total protein amount	1334	–	69	–	438	–	412	–	549	–
Ringer lactate solution
2 minutes	340	14	35	14	241	34	142	24	189	118
30 minutes	526	31	38	16	194	82	227	18	257	184
Residual proteins	287	19	9	2	14	11	94	24	101	116
Total protein amount	1153	–	82	–	449	–	463	–	547	–
Solution‐containing surfactant
2 minutes	497	326	107	3	227	20	121	23	238	162
30 minutes	552	325	112	4	174	31	143	15	245	184
Residual proteins	382	45	11	2	0	0	55	11	112	163
Total protein amount	1431	–	230	–	401	–	319	–	595	–
H2O
2 minutes	334	437	19	6	225	24	206	72	196	122
30 minutes	486	328	31	4	185	41	217	57	230	172
Residual proteins	264	5	15	4	6	8	107	43	98	108
Total protein amount	1084	–	65	–	416	–	530	–	524	–

M, mean value; SD, standard deviation.

Flow‐cell model

Using SDS extraction, blood‐soiled TOSI^® carriers showed protein values ranging from 4890 μg protein/ml to 5020 μg protein/ml. The protein elution curve of the ringer lactate solution had an elution maximum of 412 μg protein/ml after 3·5 minutes. After a purge time of 13 minutes, all proteins were purged from the test cell, and a tailing effect was noticed. In comparison, for saline solution or deionised water, the maximum concentration of the eluted proteins was obtained after 3 minutes. For water, a maximum of 463 μg protein/ml was measured in the effluent rinsing solution. For saline solution, the maximum protein concentration was lower at 301 μg protein/ml. The saline solution again showed pronounced tailing, while tailing tendencies were lower for water. The protein elution curve of the surfactant solution differed from other test solutions. The elution maximum was significantly higher (P < 0·05), with a maximum of 854 μg protein/ml after 2·5 minutes, compared to the other cleansing solutions, with less tailing as compared to saline solution or deionised water (Figure 6). Due to the construction of the test cell, it was possible to visually observe and document the cleaning process in addition to the protein determination in the effluent rinsing solution (Figure 7). The photo documentation of the purification of the TOSI^® blood test soil with water or the surfactant solution using the flow‐cell method showed an initial detachment of the test soil during the application of the surfactant solution after 1 minute. Repeating the experiments on seven consecutive days with a rinsing duration of 15 minutes, again, deionised water tended to be the most effective rinsing solution, followed by the ringer lactate solution and saline solution. Because the surfactant‐containing rinsing solution interfered with the protein measurements, no conclusion on its performance could be made using this test model (Table 2). In contrast to the findings with TOSI ^®, TOSI^® Gold test specimens did visually show that the surfactant‐containing solution had no protein‐decontaminating effect even up to 30 minutes of rinsing.

Figure 6.

Cleaning effect of the tested rinses in blood‐soiled TOSI^® test soil.

Figure 7.

Visualisation of the cleaning process after rinsing the blood‐contaminated TOSI^® test soil with water or a solution‐containing surfactant.

Table 2.

Protein concentrations after rinsing the blood‐contaminated TOSI^® test soil for 15 minutes on seven consecutive days

Protein		Deionised water	0·9% saline solution	Ringer lactate solution	Solution containing surfactant
Rinsing solution	M (μg/ml)	5937	4736	5400	7420
	SD (μg/ml)	1462	802	646	1234
	SD %	25	17	12	17
Residual protein (TOSI)	M (μg/ml)	122	59	57	109
	SD (μg/ml)	67	39	23	62
	SD %	55	66	39	57
Total protein	M (μg/ml)	6059	4795	5457	7529
	SD (μg/ml)	1422	828	654	1254
	SD %	23	17	12	17

M, mean value; SD, standard deviation.

Discussion

An important aspect of a wound‐rinsing solution is its cleaning and decontamination effect. It is important to dissolve contaminants, wound deposits, microorganisms and encrustations and remove them from the wound. Therefore, wound‐rinsing solutions, which enable optimum cleaning, would be advantageous. In our study, different wound‐cleansing solutions were tested and compared in different test models.

Chamber models

The three‐chamber method represents a feasible test model for the comparison of the cleaning performance of different rinse solutions 12. The model is simple, and measurements can be carried out without major technical effort. For the three‐chamber method, human blood plasma containing fibrinogen was used as test soiling. Fibrin test soils were prepared by the adding of calcium ions, thus inhibiting the anticoagulation. However, the measured values were not reproducible. Moreover, only placing the test specimen in the rinsing solution – which does not reflect clinical practice – simulated cleaning. The removal of a representative sample during the purification process was hardly possible because the protein could not be distributed homogeneously in the rinse solution. This was only achieved by moving the test chamber. The large fluctuations occurring between the individual measurements when using the method of Kaehn 12 were challenging and caused difficulties with the reproducibility. In addition, in the study of Kaehn 12, only the results of individual measurements had been given; therefore, the comparison with his reported findings was not possible.

Therefore, the three‐chamber model was slightly modified. Only one field of the three‐field object carrier was soiled (‘one‐chamber‐model’), and the rinsing was performed in a beaker glass. Using the one‐chamber method, the measurements still showed large fluctuations but a better reproducibility than when using the three‐chamber method. The two investigated test soils, plasma and fibrin, which differed markedly in their adhesion to surfaces. Blood plasma could be removed quickly and easily. It was removed by all investigated cleaning solutions so quickly that it was not possible to measure differences in the cleaning performances. Using BCA to measure the protein concentration, the surfactant‐containing solution showed stronger protein release behaviour than rinsing with saline solution and ringer lactate solution. On the basis of the measurement of the residual protein, however, this conclusion was found to be contradictory because the protein quantities in the effluent rinsing solution passing outward from the three‐chamber model after rinsing with the surfactant‐containing solution yielded higher protein values than those of the other rinsing solutions. Obviously, there is an influence on protein determination in the effluent rinsing solution towards false positive (too high) protein values. The ability of saline and ringer lactate solution to detach proteins was similarly weak and did not differ significantly. A slight differentiation of their efficacy was possible using the F‐slide as a test specimen, with higher efficacy of the saline solution. According to the available data, it can be estimated that the three‐chamber method according to Kaehn 12 is not suitable as an in vitro model to test wound rinsing solutions due to the poor reproducibility.

Biofilm model

In order to better investigate the cleaning performance of different rinsing solutions, an alternative test soiling with better adhesion on the test specimen was sought. Biofilms are a component of chronic wounds and adhere strongly to surfaces. Therefore, the idea of using a biofilm as a model to simulate the surface of a wound appeared obvious. However, compared to other test models, handling and implementation of the biofilm model demanded a considerably higher effort. The main challenge when working with biofilms was growing reproducible biofilms to avoid fluctuations in the biofilm quality and thus detachment of proteins. The large variance in the thickness of the biofilms led to a scattering of the measured values, and therefore, no statistical differences between measurements could be detected. However, the advantage of choosing a biofilm model was the better adhesion to surfaces and thus impeded removal in comparison to protein soils, but also the usability of biofilms for the efficacy testing of antimicrobial compounds 19, if required.

Flow‐cell model

The flow‐cell method represents a hydrodynamic system for testing the surface‐cleaning effect of various detergents 13. The main advantage of the model is that, apart from the hydrodynamic pressure, which is determined by the flow rate, any mechanical influence during cleaning is excluded. If the flow velocity is kept sufficiently low and constant, the influence of mechanics can be excluded during cleaning, and the cleaning performances of rinsing solutions can be directly compared. For the flow cell method, commercially available metal strips (TOSI^® test specimen) soiled with human blood correlate were used, which were prepared according to DIN ISO 15883. In addition, commercially available metal strips (TOSI^® gold) contaminated with denatured proteins, which exhibited a stronger adhesion, were also used as test specimens. Again, the effluent rinsing solution of the surfactant‐containing solution showed higher protein levels compared to the saline solution, ringer lactate solution and water. However, the residual protein quantity on TOSI^® test pieces was also the highest for the surfactant solution. This contradiction may be explained by an influence of the surfactant undecylenamidopropyl betain or the preservative polihexanide on the protein measurement. Saline solution, ringer lactate solution and deionised water have shown a similar cleaning behaviour. All three rinse solutions show a long right tail, which indicates a relatively slow rinsing of the test soils. The results of blood contamination (TOSI^® blood) demonstrated that the test soil was easily rinsed off from the test specimen. In order to allow a better differentiation of the cleaning performance between the various rinse solutions, the TOSI^® gold carrier loaded with denatured proteins and improved adhesive properties was used. However, these results were less significant than those of the TOSI^® blood test specimens because no protein was detectable in the effluent rinse solution, that is, protein adherence was stronger than the effect of the rinsing solution. In order to identify possibly present differences better, test soils with modified adhesion properties could be used.

Conclusion

It was found that the cleaning and decontamination effect is strongly influenced by the mechanical component during rinsing. In part, the influence of the mechanical component considerably exceeds the differences in the rinsing performances of different test solutions. This provides highly scattered results in models where the influence of the mechanics (e.g. flushing pressure) cannot be controlled, and therefore, these models are not suitable for the evaluation of the rinsing performance. In our study, the only suitable method for the comparison of wound‐rinsing solutions was the flow cell method. A further important influencing factor when comparing the rinsing performance is defined and reproducible test specimens. According to our investigations, the most suitable specimens were the commercially available TOSI^® test specimen. A further possibility is the fibrin soil that is more difficult to remove, compared to the plasma soiling. If the cultivation of biofilms could be optimised with regards to the reproducibility in the protein concentration, the biofilm represents a possible alternative with, however, a higher workload. In addition, when measurements of protein concentrations are used as a criterion for the cleansing efficacy of wound irrigation solutions, the interference of the solution with the protein measurements needs to be determined.

In the three test models, water, ringer lactate solution and saline solution did not differ in their rinsing performance. Taking into account the lack of significance concerning the rinsing performance, the physiological saline solution was superior to the ringer lactate solution in all three models, whereas water was the least effective in the biofilm model but most efficient in the flow‐cell model.