An in vitro study of the antimicrobial activity of some endodontic medicaments against Enteroccus faecalis biofilms


Dr Basil Athanassiadis
268 Ipswich Road
Annerley QLD 4103


Background:  The in vitro antimicrobial activity of a series of endodontic medicaments and their bases against biofilms of Enterococcus faecalis was investigated.

Methods:  The medicaments tested were Pulpdent paste, Ledermix paste, a 50:50 Ledermix and Pulpdent mixture, and a replica of Ledermix paste. Bases included methyl cellulose with water, polyethylene glycol (PEG), and PEG with zinc oxide, calcium chloride and the other components (inactives) that make up the Ledermix paste base. Biofilms grown on cellulose nitrate membrane filters were exposed to the medicaments for up to 5 days. The number of surviving colony forming units (CFU) was determined at days 1, 3 and 5. The results were expressed as a bacterial survival index (BSI) when compared to the unexposed control.

Results:  Pulpdent produced the greatest reduction of BSI, followed by the 50:50 mixture of Pulpdent and Ledermix pastes. Ledermix paste, its replica and the individual bases showed no significant reductions in the BSI for E. faecalis.

Conclusions:  Within the limitations of this laboratory study, calcium hydroxide containing preparations had greater potential for reducing the survival of E. faecalis in a biofilm environment.

Abbreviations and acronyms:

bacterial survival index


colony forming units


horse blood agar


polyethylene glycol


One of the most important features of a biofilm is that embedded bacteria display a marked decrease in susceptibility to antibacterial agents and to host defence systems when compared to their planktonic counterparts.1 There are several mechanisms of resistance of bacterial biofilms to antibiotics. These include the ability of the polysaccharide matrix to retard diffusion of antibiotics, quorum sensing, transfer of antibiotic and virulence traits between organisms in the biofilm, inability of the antimicrobial agent to penetrate deep into the biofilm, and the inability of the antimicrobial agent to affect bacteria which are slow growing and in a more dormant metabolic state.2–5 This was brought to light in a study in which 14 of 16 endodontically treated teeth showed residual and viable intracanal bacteria. The microbes were located in areas that were inaccessible to instruments and irrigants, and they appeared mostly in biofilms. These biofilms were not removed by instrumentation and irrigation with sodium hypochlorite in one treatment session.6

Enterococcus faecalis is relatively easily destroyed in planktonic forms in vitro, but it appears to become more resistant when it is present in an infected root canal system. This may be due to the activation of virulence factors, biofilm formation, or invasion of dentinal tubules. Although enterococci have both intrinsic and acquired resistance to many antibiotics, they are inherently more resistant to antimicrobial drugs than other clinically important Gram-positive bacteria.7 The spectrum of antimicrobial resistance of E. faecalis covers a wide range of different antibiotic classes including macrolides (such as clindamycin), beta-lactam agents (such as cephalosporins) and the aminoglycosides.8

Some of the common endodontic medicaments used in dentistry are listed in Table 1. The vehicles used in these products are methyl cellulose, polyethylene glycol (400) and glycerine. Liquids such as sterile/distilled water, saline and local anaesthetic solutions do not give bulk to the active ingredients, and are not used in medicaments, whereas viscous vehicles such as PEG 400 and glycerine improve the handling properties of these pastes and facilitate their placement into the root canal space.

Table 1.   Some common commercially available medicaments and their components. (Information obtained from material safety data sheets and references 12 and 23.)
Commercial nameCompositionpH
  1. N/A = not available.

Pulpdent35% Ca (OH)2, barium sulphate, methyl cellulose, water12.2
Ultracal XS35% Ca (OH)2, 2% barium sulphate, methyl cellulose, water12.5
ApexCal29% Ca (OH)2, 22% bismuth carbonate, glycerine, water, PEG12.4
Calen49.77% Ca (OH)2, zinc oxide, PEG 400N/A
DT Temp Dressing40–50% Ca (OH)2, methyl cellulose, water12.6
Calasept41.07% Ca (OH)2, 8.33% barium sulphate, 50.6% sterile isotonic saline12.4
Calcipulpe20% Ca (OH)2, 20% barium sulphate, carboxymethylcellulose, excipients11.8
Ledermix paste3.2% demeclocycline HCl, 1% triamcinolone acetonide, zinc oxide, PEG 400/3350, water7.6

There are no reports in the literature regarding the antimicrobial effects on biofilms of Pulpdent, Ledermix, the 50/50 combination of Pulpdent and Ledermix and their bases. The present study follows previous work using an agar diffusion model9 with the aim of evaluating the in vitro effectiveness of these commercially available medicaments and their bases on E. faecalis using a biofilm model.

Materials and methods

The endodontic medicaments evaluated were Pulpdent paste (Pulpdent Corporation of America, Watertown, MA, USA), Ledermix paste (Lederle Pharmaceuticals, Wolfsratshausen, Germany), and a 50:50 mixture of Ledermix and Pulpdent pastes. The 50:50 mixture was prepared on a volume basis. The two materials were mixed on a sterile petri dish lid using a sterile wooden stick for 20 seconds, thereby achieving a uniform mix which was no more viscous than the individual components and could be easily manipulated. A number of their vehicle base components were also tested. A preparation of water and methyl cellulose (Methocel A4M premium, Swift and Co., Rosehill, NSW) was used to replicate the base of Pulpdent paste. A mixture of polyethylene glycol (PEG) 400 and 3350 (Huntsman, Melbourne, VIC) in water was used to replicate the base in Ledermix paste. The antimicrobial action of PEG combined with the preservatives and fillers in Ledermix paste was also tested. These “inactives” comprised zinc oxide, calcium chloride, triethanolamine, sodium sulphite and Germaben IIE. A replica paste for Ledermix which lacked the corticosteroid component (triamcinolone) was prepared by combining PEG 400/3350, water, 3.2% demeclocycline HCl (Sigma, St Louis, MO, USA) and the various “inactives” mentioned above.

A single species biofilm was prepared using E. faecalis ATCC 29212. Four to five colonies of E. faecalis were subcultured from a 24-hour horse blood agar (HBA) (bioMeriuex, Brisbane, QLD) culture into 10 mL Brain Heart Infusion broth (Becton Dickinson, Sparks, MD, USA), which was then incubated at 35°C for 24 hours. The turbidity of the broth culture was then adjusted using sterile saline to give a suspension equivalent to that of a 0.5 McFarland standard. Three sterile 0.2 μm nitrocellulose membrane filter discs (Whatman, Maidstone, England) were placed onto a HBA plate. The testing employed three nitrocellulose filters per preparation in order to obtain separate results for each of the time periods (1, 3 and 5 days) with a further four filters for the control biofilm that were harvested at the same time as the test filters. This replicated the approach used in the Wilson and Thrower et al.10,11 study (with some modifications) and maintained the viability of the biofilm by providing moisture and nutrients from the agar base below. The agar medium has a maximum life of 7 days before it starts to dry out and affect the vitality of the biofilm. A 20 μL suspension of the broth was placed onto the centre of each membrane filter. After the inoculum had absorbed into the medium, the HBA plates containing the inoculated filters were then incubated for 48 hours in air at 35°C. The purity of the test strain was checked at the time of inoculation of the membranes by a subculture of 20 μL of adjusted suspension on horse blood agar.

A 0.1 gram sample of the test medicament was placed on the surface of three sterile 22 mm glass coverslips. These were then inverted over the inoculated membrane filters containing the biofilm, so that the medicament was in direct contact with the entire biofilm. Gentle pressure was applied to the coverslip so that the medicament spread to cover the biofilm, thereby ensuring each medicament was in complete contact with the entire biofilm but did not reach the edge of the filter. The plates were then incubated again at 35°C for up to 5 days. This pressure also ensured there was no air trapped between the biofilm and the medicament.

Control biofilms were also included. These were covered with a sterile coverslip which did not contain any medicament. At each time-point, one of the control biofilms was also tested. This value was then used as the baseline control CFU count for subsequent determination of the bacterial survival indices (BSI).

At designated time intervals of 1, 3 and 5 days, the coverslip and nitrocellulose discs were removed and placed into a sterile container with 10 mL of sterile distilled water. This was vortexed for one minute to disperse the bacteria, and serial dilutions prepared to 10−6 in 2 mL tryptone soya broth (bioMerieux, Brisbane, QLD) using 200 μl aliquots (i.e., a 1:10 serial dilution). A 100 μl sample of each dilution was placed on the centre of the HBA plate and spread evenly across the entire agar surface for each dilution (10−1 to 10−6 including the undiluted suspension). The plates were then incubated for 24 hours at 35°C.

The colonies were counted for each dilution, and the number of CFUs was determined using the following formula: Colonies × dilution × 100 = CFU per biofilm. The Bacterial Survival Index (BSI) was determined by taking the CFU for each test sample and dividing this by the CFU of the baseline control multiplied by 100, then the log10 was taken (i.e., CFU test divided by CFU control × 100-log10 transformed). The value thereby obtained was not the log10 of the counts, but rather a log10 of the reduction in the count of the test sample relative to the control biofilm count tested after the same length of incubation as the test. Hence the log10 reduction for untreated controls was consistently 2.0. The entire experiment was performed six times for each individual medicament and the respective base components.

The data from the individual experiments were collated and the distribution of the data sets examined using the Kolmogorov-Smirnov test. Differences between groups were analysed using the Kruskal-Wallis (non-parametric) test, with post hoc tests using Dunn’s multiple comparison test. Significance was set at the 5% level. For this study, no significant reduction was defined as being when the BSI was >1.5, a moderate reduction in BSI was between 1.0–1.5, and a significant reduction was when the BSI was <1.

In order to demonstrate the biofilm nature of bacterial growth on the filters, and assess the impact of Pulpdent on this, an image analysis method was used. Nitrocellulose discs were prepared as described above, and cultured for 1, 3 and 5 days, together with corresponding discs which had been treated with Pulpdent paste. The samples were fixed for 7 days in neutral buffered 10% formaldehyde, then processed individually as follows: each disc was stained for 60 seconds in a 1:10 dilution of a two-tone biofilm staining dye mixture (GC Plaque Check, GC Corporation, Tokyo, Japan). This material contains erythrosin and acid-fast green, and differentially stains thin biofilms red/pink and dense mature biofilms violet/blue. After staining, each disc was gently agitated for 2 minutes in distilled water to remove unbound dye, blotted for 15 seconds to remove surface water, then immediately photographed under colour corrected constant illumination from an array of light emitting diodes using a 3.5 megapixel digital camera (Coolpix 995, Nikon, Tokyo, Japan) fitted to a stereomicroscope. The camera-target distance, magnification, and camera exposure settings were identical for all samples. Each disc was processed separately so that desiccation of the discs did not occur.

Images were then processed using Adobe Photoshop software. The disc was selected from within the overall image, and histogram data collected from the entire viewable surface of the disc. This gave data from a minimum of 30 000 pixels for each disc. Histogram data for luminosity and for red, green and blue channels were collected, and means and standard deviations determined. All data sets collected from the discs showed a Gaussian distribution. Repeat measurement of one sample revealed that measurement error for the image analysis method was less than 2% of the mean. Values for test (Pulpdent-treated) and control discs at each time point were then compared using unpaired T tests with Welch correction. The same approach was used to compare treated discs with the day 0 baseline control.


Pulpdent paste resulted in a moderate reduction by day 1 and a significant reduction in the BSI by day 5, although there was a slight increase in count at day 3 (Fig 1). The reduction in the viable bacterial count (i.e., reduction in BSI) was less with the 50:50 combination of Ledermix and Pulpdent pastes after 5 days compared to Pulpdent paste used alone. Ledermix paste and its replica (no corticosteroid) showed no significant changes in BSI after 5 days. None of the bases (i.e., PEG or methyl cellulose) exerted any significant inhibitory effect, thus any reduction in the BSI values can be attributed solely to the active ingredients. Demeclocycline HCl present in both Ledermix paste and its replica, did not give reductions in BSI. In contrast, calcium hydroxide in Pulpdent and in the 50:50 combination of Pulpdent and Ledermix gave significant reductions in bacterial survival. There was no additive effect of the 50:50 Ledermix and Pulpdent combination in terms of reduction in BSI when compared to Pulpdent paste alone.

Figure 1.

 Effect of various preparations on the bacterial survival index (BSI) of Enterococcus faecalis in a mono-species biofilm model. The BSI, expressed on a logarithmic scale, is shown on the vertical axis. Control data can be seen in Table 2. Only the calcium hydroxide preparations of Pulpdent and Pulpdent mixed with Ledermix were highly effective at suppressing the survival of E. faecalis.

The bacteriostatic action of the tetracycline-based medicaments (Ledermix and its replica) was demonstrated by an initial reduction in BSI at day 1 followed by an incremental increase in BSI at 3 and 5 days. This was in contrast to the two calcium hydroxide containing preparations, where the BSI was lower at day 5 than at day 1. Averaging the percentage kill values on day 5 produced the following results: Pulpdent 91.04%, Ledermix/Pulpdent 74.70%, and Ledermix 18.28% kill.

According to Dunn’s multiple comparison test, there was a significant difference (p <0.01) for Pulpdent and the 50:50 mixture of Ledermix and Pulpdent pastes compared to the bases. Thus, of the commercial products tested, Pulpdent paste gave the greatest overall reduction of the BSI of viable E. faecalis cells in this biofilm model.

Table 2 shows the data for the controls which are the log CFU of the surviving E. faecalis bacteria on the biofilm induced membrane filters at day 1, 3 and 5. These data were constant across most time periods and amongst different medicaments and bases tested. As the results were compared to the untreated controls, the changes that occurred could be attributed directly to the effect of the medicaments/bases against the biofilm tested.

Table 2.   Colony forming units (CFU) log transformed for the controls
MedicamentDay 1Day 3Day 5
  1. Control data shown are means and standard deviation from 2 to 5 repeats for each of the medicaments/bases tested.

PEG + water9.29 (0.10)9.38 (0.22)9.36 (0.10)
PEG + inactives9.24 (0.07)9.10 (0.41)9.20 (0.20)
Methyl Cellulose + water9.37 (0.09)9.04 (0.79)9.21 (0.43)
Pulpdent9.29 (0.09)9.33 (0.11)9.33 (0.06)
Pulpdent + Ledermix (50/50)9.24 (0.20)9.35 (0.37)9.25 (0.32)
Ledermix paste9.23 (0.13)8.95 (0.44)9.01 (0.20)
Ledermix replica (no steroid)9.52 (0.11)9.38 (0.12)9.38 (0.12)

The image analysis data revealed that Pulpdent-treated discs showed a different pattern of dye staining, with a reduction in all four parameters (luminosity, red, green and blue channels) compared to the baseline control (day 0). There was also a reduction when compared to the relevant time-matched controls across the five-day experimental period, signifying a reduced level of bacterial growth overall in the treated samples (Fig 2). This reduction was significant at all time periods across 1–5 days (p <0.001).

Figure 2.

 Intensity data for control and treatment discs derived from image analysis. Bars show means and standard deviations. Values for treated discs are less than the baseline day 0 control and the corresponding untreated controls across days 1–5.


This study provides insight into the physical properties of the endodontic medicaments investigated and their bases. Aqueous vehicles such as water, saline and methylcellulose promote rapid ion liberation, whilst viscous organic vehicles such as glycerine and PEG release calcium and hydroxyl ions much more slowly. This difference may explain why the latter may remain in the canal for 2–4 months.12–14

The present study used a biofilm model to assess the antimicrobial effects of different endodontic medicaments. Knowledge of biofilm formation in root canals is limited, although it is clear that biofilms may be present on the dentine walls of root canals and also on the external root surfaces of the apical portion of teeth with infected root canal systems.2 Biofilms produced in the laboratory setting may be less adherent and less virulent than those found in nature, and this should be taken into account when considering in vitro and in vivo correlations.10 The membrane filter-based model offers a simple means of assessing the susceptibility of oral bacterial biofilms to antimicrobial agents.11 This model does not account for anatomical variations seen in the root canal environment, polymicrobial nature of biofilms, the buffering effects of dentine, the effects of irrigation and the anaerobic nature of the root canal. The agar diffusion model looks at the ability of medicaments to inhibit the growth of micro-organisms in a simple cellular lawn by measuring the size of the inhibition zone whilst in the biofilm model, the medicament is placed in direct contact with the biofilm and quantitative counts are performed with formal estimation of the ability of medicaments to inhibit growth. This latter method may have value as a rapid primary screening test to assess the antimicrobial effect against biofilms.

In a previous study using a biofilm model, Distel et al.15 observed biofilms in the root canals of single-rooted human teeth that were left empty for two days with no medication. In contrast, canals filled with Pulpdent paste were colonized only after 77 days. The inability of antibiotics to penetrate biofilms was shown by Lima et al.16 in a study where 2% clindamycin allowed the overgrowth of E. faecalis biofilms in a cellulose-nitrate membrane filter model, whereas combinations of 2% clindamycin and 10% metronidazole gave very small reductions in viable bacterial cell counts. Takahashi et al.17 showed that tetracycline HCl, minocycline HCl, doxycycline hyclate and ofloxacin had no effect on three of four P. intermedia strains tested in the same biofilm model. Norrington et al.18 showed that five antibiotics (ampicillin, doxycycline, clindamycin, azithromycin and metronidazole) when used at their minimum inhibitory concentrations were entirely ineffective against endodontic bacterial biofilms after 8 days of growth, and concluded that higher concentrations of antibiotics would be required to demonstrate an effect. They confirmed that endodontic bacteria form mature biofilms that are resistant to clinically achievable levels of antibiotic therapy.18 Although BioPure MTAD (which contains doxycycline hyclate) is used as an irrigant, Dunavant et al.19 and Clegg et al.20 have shown that it had limited effects against E. faecalis biofilms.

A key point when assessing earlier work is the nature of the calcium hydroxide product tested. In a study by Chai et al.,21 the authors tested calcium hydroxide as a saturated solution, a situation which is not representative of what is placed in the root canal in clinical practice. The antibiotics tested in that study were injectable liquid forms with each having different vehicles, including benzyl alcohol, propylene glycol, lidocaine and other vehicles. Using an E. faecalis membrane filter model in which discs were incubated for 48 hours and then exposed to antimicrobial agents for one hour, they showed that erythromycin, oxytetracycline and calcium hydroxide (a saturated solution) were 100% effective in suppressing subsequent growth of E. faecalis.21 A saturated calcium hydroxide solution was also used by Abdullah et al.22 which gave a six-log reduction in bacterial counts of E. faecalis after 60 minutes of contact.

The biofilm model used in this study has produced different results to those obtained in earlier work using an agar diffusion model, which showed a similar zone of inhibition for E. faecalis with each of the same commercial products tested in this study.9 An explanation for the greater effectiveness of the calcium hydroxide containing pastes such as Pulpdent paste in the biofilm model used in the present study may be the greater penetration of hydroxyl ions into the biofilm compared to the antibiotics which have larger and more complex molecular structures. Also, antibiotics may not be as effective as biocides due to the various mechanisms of bacterial resistance in biofilms, as outlined earlier.

The variations in BSI values (i.e., standard deviations) between individual biofilms tested under the same conditions may be explained by the heterogeneity of biofilm composition and their growth dynamics. In addition, the methods used in this study measure only the survival/death of E. faecalis. Tetracycline-based materials such as Ledermix are bacteriostatic rather than bactericidal (i.e., they only inhibit growth and do not necessarily kill the organisms). This may explain why the BSI increases at day 3 and 5 for these medicaments, whilst there is a further reduction in BSI at day 5 for the calcium hydroxide containing medicaments. The antimicrobial activity of tetracycline-based materials may not be revealed fully using this biofilm model. It is not clear why the day 3 BSI for Pulpdent is higher than that of either day 1 or day 5, but this pattern may reflect the complex nature of biofilms and their characteristics.

In summary, this study has shown that Pulpdent paste had the greatest overall reduction of the BSI of viable E. faecalis cells in a mono-species biofilm model using a cellulose-nitrate disc model. The 50:50 combination of Ledermix and Pulpdent pastes gave a moderate reduction in the BSI and there was no synergistic effect, whilst Ledermix paste, its replica and the bases resulted in no significant changes in the BSI.