Methodology.
The hypothesis was tested ex vivo that Nd:YAG laser irradiation has a bactericidal effect on endodontic pathogens inoculated in root canals. Resultant colony-forming unit counts were associated with observations of bacterial cell structural changes using conventional scanning electron microscopy (CSEM) and environmental scanning electron microscopy (ESEM) on inoculated dentine surfaces, following indirect and direct Nd:YAG laser irradiation, respectively.

Results.
The Nd:YAG laser irradiation (1.5 W, 15 Hz, four times for 5 s) of Enterococcus faecalis inoculated canals resulted in a significant reduction (P < 0.05, Wilcoxon signed rank test) of the bacterial load, meaning a 99.7% kill, but no sterilization. The CSEM procedure verified that the extent of radiation damage was in line with the total amount of laser energy applied. After 2 h of incubation and three cycles of indirect laser treatment (i.e. through a 1-mm-thick dentine disc), no morphologically intact bacteria of Actinomyces naeslundii or Streptococcus anginosus were discernible. However, when micro-colonies of S. anginosus and specially biofilms of E. faecalis were present after 2 days, the in situ experiment using ESEM and direct laser treatment showed that bacterial eradication was reduced in deep layers.

Conclusions.
The Nd:YAG laser irradiation is not an alternative but a possible supplement to existing protocols for canal disinfection as the properties of laser light may allow a bactericidal effect beyond 1 mm of dentine. Endodontic pathogens that grow as biofilms, however, are difficult to eradicate even upon direct laser exposure.

Introduction.
Antisepsis is the essence of root canal treatment in teeth with apical periodontitis (Kakehashi et al. 1965). As a result, the use of mechanical instrumentation and sodium hypochlorite (NaOCl) is routine in clinical practice. When irrigating with NaOCl, fine gauge needles can be used to facilitate deep penetration but agitation using small size instruments or ultrasonics (Lee et al. 2004b) may increase the potential to kill microorganisms (Huque et al. 1998). Smear layer removal from the canal wall by citric acid or ethylenediaminetetraacetic acid (EDTA) has been recommended (Scelza et al. 2004), although there is no clear evidence that this procedure enhances disinfection or treatment outcome (Bystro¨m & Sundqvist 1985, Ørstavik & Haapasalo 1990).
Unfortunately, careful use of stainless steel hand and/ or nickel–titanium rotary files with NaOCl in one visit cannot render all systems bacteria-free (Sjo¨gren et al. 1997, Shuping et al. 2000, Nair et al. 2005). Whilst microbial killing may continue after filling because of the antibacterial properties of sealer and/or guttapercha (Saleh et al. 2004) or by blocking access to nutrients (Sundqvist & Figdor 1998), filling root canals having a positive culture will, in general, adversely affect treatment outcome (Sjo¨gren et al. 1997).
Some investigators advocate the use of calcium hydroxide [Ca(OH)2] as an intracanal medicament in a multiple-visit approach (Bystro¨m et al. 1985, Shuping et al. 2000). The environment within the system, however, is such that delivering the medicament and maintaining a high pH homogeneously is a challenge (Nerwich et al. 1993). Careful packing of the calcium hydroxide is essential because if not in direct contact with its target it is unreliable at best (Bystro¨m et al. 1985). In addition, root dentine appears to reduce the activity of Ca(OH)2 by buffering the local pH (Haapasalo et al. 2000).
Despite the use of Ca(OH)2, certain microbial species in a limited group of cases do survive and can be held responsible for persistent infections (Ørstavik & Haapasalo 1990, Gomes et al. 1996, Sundqvist et al. 1998, Chavez De Paz et al. 2003). Enterococcus faecalis, for instance, is able to withstand a pH of 11.1 due to a functioning proton-pump (Evans et al. 2002). In addition, pure cultures of this species can form protective biofilms in Ca(OH)2-medicated canals (Distel et al. 2001) and have the ability to invade dentinal tubules (Love 2001, Peters et al. 2001). Under starved conditions, E. faecalis shows resistance to NaOCl (LaPlace et al. 1997), whilst the upregulation of stress-induced proteins has been shown to be important for cell survival (Hartke et al. 1998).
Dealing with persistent infections remains a challenge even though various strategies exist. One such procedure is the use of chlorhexidine (CHX) or iodine potassium iodide (IKI) alone or in combination with Ca(OH)2 (Safavi et al. 1990, Haapasalo et al. 2000, Gomes et al. 2003, Baker et al. 2004, Siren et al. 2004). Alternatively, a larger preparation and ‘deep shape’ may facilitate flow of the irrigants (Albrecht et al. 2004, Lee et al. 2004a). In this respect, the diameter of the apical preparation appears to be less important for debris removal and bacterial removal when a certain taper is achieved (Card et al. 2002, Coldero et al. 2002). One may further assume that improved shaping facilitates packing of the medicament, especially in systems with complex anatomy. Overall, the treatment of choice should be effective, reproducible and minimally invasive.
A controversial way to manage pathogens is to use lasers. The efficacy of neodymium:yttrium–aluminum– garnet (Nd:YAG) lasers for photo-thermal disinfection has been investigated and laser application was found to be safe and have potential (Klinke et al. 1997, Ramskold et al. 1997, Moritz et al. 1999, Schoop et al. 2004). Unfortunately, Nd:YAG laser irradiation could not render all systems bacteria-free with the result that no superior effect when compared with NaOCl irrigation occurs (Hardee et al. 1994, Moshonov et al. 1995). Given the high costs of lasers, clarity should be provided before advocating their use as an alternative or even a supplement to existing protocols.
This investigation ex vivo was performed to better define the role of Nd:YAG lasers in minimally invasive root canal disinfection (i.e. without further dentine removal). The hypothesis was tested that Nd:YAG laser irradiation has a bactericidal effect on endodontic pathogens inoculated in root canals. Resultant colony- forming unit (CFU) counts were associated with observations of bacterial cell structural changes using conventional scanning electron microscopy (CSEM) and environmental scanning electron microscopy (ESEM) on inoculated dentine surfaces, following indirect and direct Nd:YAG laser irradiation, respectively. With ESEM, samples can be imaged without prior dehydration and conductive coating, thus allowing real time observation and testing of endodontic treatment modalities in situ (Bergmans et al. 2005).

Materials and methods.
Microbiological analysis.
Eight maxillary incisor teeth were stored (0.5% solution of chloramine in water) at 4 _C. At the time of use, root surfaces were cleaned using an ultrasonic scaler (P5 Booster, Satelec, Merignac, France) and treatment was initiated by creating access and by removing pulpal remnants. After pre-flaring with Gates Glidden burs sizes 1–4 (Dentsply Maillefer, Ballaigues, Switzerland), each canal was instrumented using the GT rotary system (Dentsply Maillefer) in a crown-down sequence up to a size 30, 0.10 taper instrument at 1 mm from the apical foramen. Throughout instrumentation, irrigation with a 2.5% NaOCl solution was performed using a 27-gauge needle (Monoject, Sherwood Medical, St Louis, MO, USA) and patency was assured by placing a size 08 K-file (Dentsply Maillefer) 1 mm beyond the apical constriction. Apical foramina were sealed with a restorative material (Z100, 3M ESPE, Seefeld, Germany) and root canals and outer surfaces were irrigated with 17% EDTA for 2 min, followed by tap water. The prepared teeth were mounted in bijou bottles (Fig. 1a) and the assembled entities were autoclaved (134 _C for 15 min) to obtain sterilized systems (procedure checked in sampling-based pilot study and by negative controls). The bijou bottles were then filled under strict asepsis with sterile Brain Heart Infusion (BHI) broth (37 g L)1, Oxoid Ltd, Basingstoke, UK) until the available root was surrounded (method according to Seal et al. 2002).


Figure 1. Sample preparation (a) for microbiological analysis: Prepared maxillary incisor teeth were mounted in bijou bottles; (b) for CSEM: Molar crowns were sectioned in the occlusal plane near the pulp expensions. Each crown was sectioned a second and third time coronally from the first cut, resulting in two dentine discs of 1mm thickness; (c) for ESEM: Molar crowns were prepared in such a way that mounting them in a specimen stub was possible with a flat surface positioned horizontally. This surface was cut in the occlusal plane near the pulp extensions.

A bacterial suspension of E. faecalis (LMG 7937) from the Belgian Coordinated Collections of Microorganisms (BCCMTM, Gent, Belgium) and standardized (4 • 108 CFU mL)1) in BHI broth was inoculated into six of the prepared root canals using sterile syringes (Monoject). Next, the samples were incubated for 2 days (Concept 300 Anaerobic Workstation, Ruskin Technology, West Bradford, UK) under anaerobic conditions (80% N2, 10% CO2 and 10% H2) at 37 _C in isolated boxes. The two remaining samples were uninfected (sterile BHI broth, negative controls).
After incubation, residual medium within the root canals was removed with sterile paper points and replaced by 0.9% sterile saline. Afterwards, the six infected teeth were randomly assigned to either a positive control (no laser treatment, n ј 3) or a laser irradiation (n ј 3) group.
In the experimental group, laser irradiation without water spray, air cooling or photosensitizing dye was performed in wet root canals with a Nd:YAG laser (Smarty A10; DEKA, Firenze, Italy) (wavelength of 1064 nm). Standardized settings were: power output (1.5 W), energy (100 mJ) and length (150 ls) of the pulse, and frequency (15 Hz). Light was transferred by means of a 300-lm-thin flexible fibre and the actual output was controlled with a power meter. The target beam was generated by a He/Ne laser (632.8 nm, 1 mW). Throughout laser treatment, the fibre tip was applied with a spiral movement starting 1 mm above the apex and then moving coronally, four times for 5 s, interleaved with 20-s recovery intervals. For the negative controls, the same procedure was performed without activating the laser.
After treatment, the liquid contents of the root canals of all groups were carefully absorbed with sterile paper points without intentionally touching the walls. The root canals were then filled with Reduced Transfer Fluid (RTF) and gently filed in a circumferential way using sterile, size 25 K-files (Dentsply Maillefer) to working length for 20 s. Next, the contents were again carefully absorbed with sterile paper points, transferred to 2 mL of RTF, and immediately processed. After vortexing for 30 s, the contents of the test tubes were serially diluted in log steps (20 lL in Eppendorf tube with 180 lL saline, five times, thus diluted to 10)5) and duplicate samples of 50 lL were applied to Blood Agar (BA) plates (Blood Agar Base II; Oxoid Ltd, Basingstoke, UK), which were supplemented with haemin (5 mg mL)1), menadione (1 mg mL)1), 5% sterile horse blood, and 0.8% (w/v) Bacto Agar (Difco Laboratories, Detroid, MI, USA), and incubated under anaerobic conditions at 37 _C. After 3–7 days, contamination by other species was ruled out and colonies of E. faecalis were counted using a stereomicroscope at 16• magnification (Zeiss, Oberkochen, Germany) and recorded as number of CFU mL)1.
To test reproducibility, the teeth of the experimental and positive control group were re-used being assigned to the other group (cross-sectional design). For this purpose, the content of each root canal and bottle was removed, and the systems were returned to their baseline state (irrigated with tap water for 2 min, reassembled as a tooth-lid-bottle entity and autoclaved). Statistical analysis was performed using the Wilcoxon signed rank test (nonparametric paired t-test).