Skin sampling-validation of a pad method and comparison with commonly used methods A. Hambraeus, Department *M&lycke
and W. Whyte?
of Clinical Bacteriology, University Hospital, Uppsala, Sweden, Health Care AB, Sweden, and TBuilding Services Research Unit, University of Glasgow, Scotland Accepted
Two types of pad were tested for sampling bacteria from the skin. One pad was made of 85% viscose and a 15% mixture of polyester and polyamide fibres, the other was made of polyvinyl-alcohol foam. The efficiency was calculated using double samples and 8 consecutive samples. The two pads were equally efficient and more efficient when moist. Their efficiency was compared with that of the bud swab, the Rodac plate and the scrub cylinder method. Using the double sample method for calculation, the scrub cylinder method and the pads were equally efficient, around 50%. As calculated from consecutive samples the efficiency of the pads was about 45%, the efficiency of the cylinder scrub method 29%, and that of the bud swab and Rodac plate 16% and 5% respectively. The low efficiency of the Rodac plate may be partly explained by the fact that microcolonies and not bacterial cells are sampled with this method.
Introduction Several methods have been used for sampling bacteria from skin. The bud swab and agar contact methods (Hall & Hartnett, 1964; Updegraff, 1967) are commonly used for skin sampling. The most accurate skin sampling method for experimental work is probably the scrub-cylinder method described by Williamson & Kligman (1965). However, this is elaborate to use. A velvet pad imprint method has been used by Raahave for wound of this method was described by sampling (1975a,b). A modification Benediktsdottir & Hambraeus (1983). They used a pad made from 85% viscose and 15% polyester, and polyamide fibre (VPP pad). This pad not only takes an imprint but also absorbs liquid and was shown to sample 62% Correspondence Uppsala, Box
Anna Hambrseus, S-751 22 Uppsala,
0 1990 The Hospital
A. Hambraeus et al.
of Stuphylococcus aureus inoculated onto rabbit flesh. It was also found to be more efficient than the washout method when used for sampling bacteria from operating wounds in hip surgery. The pad method is readily handled by nurses and bacteriologically untrained staff and it was thought it could be used for sampling skin as well as wounds. As the VPP pad sheds fibres, a pad made from polyvinyl alcohol foam (PVA) has also been developed. The aim of the present study is to evaluate the efficiency of the pad method and to compare it with more commonly used skin sampling methods. Materials
Description of pads VPP pad. This is a non-woven
material made of 85% viscose and a 15% mixture of polyester and polyamide fibres (Kg Henry Hulthen, GMBH and Company, Hamburg, West Germany). PVA pad. This is made from 1 mm polyvinyl alcohol foam available world wide from suppliers of room cleaning materials.
The autoclavable non-absorbing plastic applicator currently used to hold the pads when sampling skin is shown in Figure 1. The top surface which holds the pad measures 5 x 5 cm. When using VPP pads ‘Velcro’ strips are glued to the top to hold the pads. For PVA pads thin strips of double-sided adhesive tape are used (sets consisting of an applicator PVA pad are available from Mijlnlycke Health Care AB, Molnlycke, Sweden.) In some early experiments a 5 X 5 x 2 cm piece of hard wood with a small wooden handle was used. This was not very satisfactory because the wood, if not premoistened, absorbs the water from the pads during storage. There is also a risk that wood or paint can be bactericidal.
Moistening the pads VPP pad. The standard procedure
was to stick the VPP pad to the applicator, dip the pad into a solution of 2% polysorbate 80 (Tween 80) with 0.3% lecithin in phosphate buffer solution pH 7.3 (PBS) and squeeze out the excess solution. To retain the moisture, the pads were then autoclaved in a sealed polycarbonate container in a downward displacement autoclave. After sterilization the pads contained about 1.5 g of liquid. In the experiments on the effect of different contents of liquid the pads were autoclaved dry and a known amount of liquid added aseptically. PVA pad. The PVA pads were either autoclaved in the dry state and aseptically attached to the applicator or assembled and sterilized by irradiation. Before use they were dipped into a sterile solution and squeezed against an absorbing surface to remove excess liquid. They then contained about 0.3 ml of liquid.
of skin sampling
Skin sampling An 11 X 15 cm sterile cardboard frame with a 5 X 10 cm aperture was placed over the skin. The pad was rubbed back and forth ten times, that is five double passes against the skin in the opening.
bacteria from the pads
After use the pads were ‘stomached’ for 1 min in a plastic bag with 50 ml phosphate buffer (pH 7.3) in a Stomacher Lab-Blender 400 (Seward Laboratories, London). Viable counts were made from the liquid by spreading 0.5 ml onto the surface of a blood agar plate and 0.5 ml onto a plate containing anaerobic medium.
with the Cylinder-Scrzib
A sterile cylinder with an internal diameter of 2.6 cm was placed on the skin. One ml of phosphate buffer (pH 7.3) containing 0.1% Triton X-100 was pipetted into the cylinder and rubbed over the skin with a rounded glass rod for 30 seconds. The 1 ml was removed and 0.5 ml spread over the surface of a blood agar plate and 0.5 ml over a plate containing anaerobic medium.
with a swab
A sterile cotton bud swab was dipped into 2% polysorbate 80 (Tween 80) with 0*3% lecithin in PBS and rubbed over an area of 5 X 5 cm (delineated
by a thin cardboard frame). The rubbing was carried out using about 20 lateral passes. The swab was then rubbed over two agar plates using twenty passes on each plate. One set of plates was incubated aerobically, the other anaerobically.
with Rodac plates
Blood agar with 0*5% polysorbate 80 (Tween 80) was used. The plates were rolled once over the skin to ensure that there was no multiple sampling of bacteria existing as microcolonies on the skin. One set of plates was incubated aerobically, the other anaerobically.
Bacteriology Sampling was carried out in the laboratory and all samples were processed without delay. For aerobic cultivation Columbia Blood Agar Base with 5% horse blood was used, and for anaerobic cultivation Brain Heart Infusion Broth with yeast extract, vitamin K and haemin was used (Holdernan et al., 1977). The plates were incubated aerobically at 37°C for 48 h or anaerobically for 4-6 days. No species identification was made. Anaerobic counts represented facultative as well as strict anaerobes.
of eficiency and consistency of sampling methods. were made according to Whyte et al. (1989).
Statistical analysis Double samples. Two consecutive samples were made from the same site. The efficiency
in percent was calculated using the equation: E=(l
where S, is the number of bacteria recovered in the first sample and S, the number of bacteria recovered in the second sample. Sequential samples. Eight sequential samples were made from the same site. A regression equation was calculated from the sample number less one number and the log of the bacterial counts. Efficiency is given from the slope of the line and the consistency of the method from the regression coefficient (R*).
Experimental design The eflect of degree of moisture
on skin sampling
Pairs of dry pads were compared with pairs of pads moistened with 2% Tween 80 and 0.3% lecithin in phosphate buffer. The moistened VPP pads contained 1, 2, 3 and 4 g and the moistened PVA pads contained 1 and 2 g. The largest amounts of liquid which the VPP and PVA pads could absorb was 4 and 2 g, respectively.
of skin sampling
Skin samples were taken from arms of 11 volunteers. Their left and right arms were divided into three areas avoiding the wrist and intercubital fossi where high counts could be expected. One arm was sampled only once on a single day and the sites were randomly chosen so that each area was sampled only once with a pad with a certain wetness. With VPP pads six sampling rounds with all varieties of moisture were made on one volunteer, five on the second volunteer and three on the third volunteer. With PVA pads two sampling rounds were made on each of eight volunteers. The mean yield of bacteria for each person at each ‘wetness’ was used to calculate the percentage increase in yield of bacteria with wetness for that person. Eficiency and consistency of the pad sampling method-comparison with other skin sampling methods The following methods were compared: Cylinder Scrub method, Rodac dishes, swabbing, VPP pads, PVA pads. Eficiency calculated using double-samples. The skin sampling methods were carried out on both arms of four volunteers. The six sites were selected as described earlier and the sampling randomized so that no double sample was made twice on the same site by the same method. Fourteen sets of experiments were carried out. Aerobic and anaerobic counts were obtained. The efficiency was calculated according to equation (i) as described earlier. (b) Eficiency and consistency calculated using multiple samples. A site on a volunteer’s arm was sampled sequentially eight times by one of the five methods. Fourteen experiments were performed with each method on three volunteers. The efficiency and consistency of the various methods was calculated as described earlier. Results
Moist vs dry pads Table I shows the mean increase of bacteria, in percent, for the two types of pads. The recovery of bacteria was at least ten times higher with moist pads compared to dry pads but the degree of moisture did not appear to much affect the recovery. Method eficiency using double samples Table II gives the average of the bacterial counts from the first sample. The variations in skin counts between the test persons were large but the ranking list was much the same for all the methods. The highest counts were always found with the pad methods and the lowest with swab and Rodac plates. The highest mean value was 332 bacteria from the anaerobic count obtained when sampling with the VPP pad and the lowest mean value was 2.9 from the anaerobic count when swab sampling. Statistical calculation using 2-factor analysis of variance showed that there was no statistical difference between the Rodac and the Swab method or between the two pad methods.
I. Percentage increase in yield of bacteria with increasing liquid in pad compared with dry pad Comparison with Mean % increase
Amount of liquid in pad k)
dry pad [range]
IO.5 x 102 [(75-l 5) x 107
23.6 x lo2 [(126-41~1) x 107
9.9 x 102 [(l+34.6) x 102]
10.6 x IO* [(1.7-45*9)x lo*]
9-6 x lo* [(7-11.7)X 107
18.2 x lo2 [(16.3-21.7)x 102]
5.4 x 102 [( -04-13~5) x 107
6.4 x lo2 [(O-8-24.1) x 102]
7.8 x lo2 [(2-O-14.8) x lo*]
14.9 x 102 [(7*2-27.6)x lo*]
6.3 x lo2 [(52-8.4)x 102]
18.2 x lo* [(13.8-25.9)x
Yield of bacteria by five different sampling methods Sampling VPP
Test person 1 2 :
Overall mean values
No. experiments 3 4 43
190 45 726 90.6
310 29.6 12.5 864
350 294 416 51
680 96 244 76.6
22.3 13.6 158 34.8
71.9 9.2 63.3 36
1.9 1.8 111.3
4.1 1.4 4.8 1.4
3.8 2.9 192.0
11.7 1.4 1:.: *
a VPP and PVA methods, no significant from all other methods (P