Introduction

Safe and effective irrigation is of central importance to successful root canal treatment. It fulÞlls several important mechanical, chemical, and (micro)biological functions. Instrumentation of the root canal system must always be supported by irrigation to remove pulp tissue remnants and dentin debris. Without irrigation, accumulation of this debris causes instruments to rapidly become ineffective. Several irrigating solutions also have antimicrobial activity against bacteria and yeasts. A bigger challenge for irrigation may be the areas untouched by the Þles, such as Þns, isthmuses, and large lateral canals. Also large areas in oval and ßat canals may remain untouched despite careful instrumentation. These areas contain tissue remnants and bioÞlms which only can be removed by chemical means using irrigation. In order to simulate this in vivo situation, a variety of in vitro bioÞlm models are currently used in endodontic research, for example, to study how irrigation and instrumentation can kill bioÞlm bacteria and remove these bioÞlms. Factors that remain a challenge with irrigants include poor penetration, limited tissue-dissolving ability, and exchange in the highly complex root canal anatomy. Optimal irrigation is based on research using reliable, reproducible, and standardized irrigation models that closely replicate in vivo scenarios in order to predict safe and effective irrigation. New developments such as computational ßuid dynamic models help to interpret and better explain the outcomes of ex vivo, microbiological, and clinical studies and help with the design of new strategies. This article presents an overview of the various factors that need to be considered when developing models to study the effect of irrigation on endodontic bioÞlms, tissue remnants, and debris removal. We attempt to explain how differences in experimental methods may affect the reported behavior, as well as to provide cutting-edge information on recent developments.

system. Although this could be accomplished by optimal chemomechanical instrumentation [1], it is difÞcult to predictably reach this goal [2Ð4] because of the complex structure of the root canal system and the resistance of bioÞlms [5Ð7]. Instrumentation of the root canal system must always be supported by effective irrigation. The efÞcacy of an irrigation system is dependent on its ability to deliver the irrigant to the apical and uninstrumented regions of the canal space, to clear the debris from the canals [8Ð12], to dissolve necrotic tissue and bioÞlm, and to kill planktonic and bioÞlm microorganisms. Although many new developments have taken place with introduction of new irrigating solutions and equipment, there is currently no solution or method that predictably results in completely cleaned root canals [13Ð24].

In 1981, Bystršm and Sundqvist [25] reported that mechanical instrumentation and saline irrigation greatly reduced bacterial numbers in the infected root canal. However, in ca. 50 % of the cases, bacteria could still be detected in the canals after four appointments. Nevertheless, mechanical instrumentation has been considered one of the most important phases in endodontic treatment. In a study by Dalton et al. [26], the root canals were prepared, irrigated with saline solution, and sampled for microbial growth from the canals before, during, and after instrumentation. The results showed that while progressive Þling by both rotary and stainless steel hand instrumentation reduced the number of bacteria, none of the techniques resulted in germ-free canals. Similar results were reported also by Siqueira et al. [27] who reported a 90 % reduction in bacteria counts by instrumentation combined with saline irrigation. In another study the authors reported that 1Ð5 % sodium hypochlorite (NaOCl) solutions were signiÞcantly more effective than saline in reducing bacterial counts in the root canal [28].

The goal of endodontic therapy is the removal of all vital or necrotic tissue, microorganisms, and microbial by-products from the root canal

Traditionally, CFU counts of bacteria have been used as the gold standard method for evaluating the effectiveness of disinfection. Different

experimental designs have been employed, including (1) direct contact tests in vitro, (2) ex vivo studies using contaminated root canals in extracted teeth, and (3) in vivo studies.

In Vitro: Direct Contact Tests

A traditional way of measuring the antimicrobial effectiveness of endodontic irrigants and disinfecting solutions has been with direct contact tests in test tubes. Bacteria in known concentrations (CFU/mL) are incubated for different time periods in disinfecting solutions such as NaOCl and chlorhexidine (CHX) of various concentrations, sampled, diluted, and cultured on solid media, for example, which allows for counting the CFUs after a period of growth [29Ð31]. Despite the seemingly simple design, the results from different studies have shown considerable variation. There are several reasons for the differences in different studies. The two main reasons are non-standardized exposure conditions and the use of microbial cultures which are at different, often unidentiÞed growth phases. In several studies, bacteria were exposed to the disinfectants while still in their growth medium [29, 30]. This invites several confounding factors, which can greatly impact the results. The culture medium contains a variety of compounds that may inhibit the activity of the antibacterial substances [32Ð 34]. In addition, if the microbes have been grown in a liquid culture for some time, the pH of the broth drops, often even several pH steps. The activity of many disinfecting agents such as calcium hydroxide and NaOCl is dependent on the pH. When the experimental conditions are properly standardized and reported, the results can be expected to be more constant. Nevertheless, direct contact tests with planktonic bacteria cannot replicate the in vivo conditions and the results must be interpreted with great caution. However, a study comparing the effectiveness of disinfecting agents against bacteria in simple in vitro killing studies with planktonic bacteria to results obtained using killing in bioÞlms indicated that the planktonic killing tests can predict the ranking of the effectiveness of the same agents in bioÞlms [35]. However, planktonic studies give much too

optimistic picture of the sensitivity of root canal bacteria to these agents. Therefore, bioÞlms are today recommended instead of planktonic bacteria for direct contact tests [35].

The agar diffusion test and CFU counting method have traditionally been used to measure the effectiveness of endodontic disinfecting solutions [28Ð31]. Unfortunately, both of these methods have considerable shortcomings. The use of the agar diffusion method to test the antimicrobial activity of endodontic materials is not based on accepted standardization of the methods. Chemical interactions between the media and the disinfecting agents are not known. Furthermore, there are no studies that would assist in drawing conclusions from the size of the zones of inhibition to the effect of the same agent in vivo in the root canal. Therefore, the information obtained from agar diffusion studies does not reliably reßect the in vitro or in vivo antimicrobial activity of endodontic antimicrobial agents and should not be used anymore [36]. However, this should not be confused with agar diffusion tests that are used to determine the effectiveness of systemic antibiotics against speciÞc bacteria, which is still a valid method for that purpose.

In Vitro/Ex Vivo: Use

of Extracted Teeth

The use of teeth or dentin blocks in in vitro and ex vivo studies of endodontic disinfection is an effort to bring the experimental conditions closer to the in vivo reality of the root canal than direct contact tests with planktonic bacteria. Often a single species, such as E. faecalis, or a mixed bacterial ßora obtained from an endodontic infection or from the oral cavity is incubated in the root canal space for 1 day to several weeks [37Ð 44]. After the incubation period, different kinds of treatment procedures are completed, and microbiological samples are taken for culture and CFU counting [37, 38, 40, 43]. Although useful information has been obtained from these studies, the dentin block/extracted tooth model has also weaknesses. In several studies, the extent of bacterial growth on the root canal wall and in the dentin canals was not veriÞed, which leaves some

room for error. In addition, the time of incubation with the bacteria and frequency of nutrient exchange show great variation [40, 43, 45, 46]. Within the Þrst hours, the bacteria are likely to be mostly planktonic and in either the exponential or stationary growth phase; bioÞlm formation is in its early stages. Portenier et al. [47] showed that planktonic bacteria in the starvation phase can be 1,000 times more resistant to disinfecting agents than the same bacteria in the exponential or stationary phase. Another key factor affecting bacterial sensitivity is bioÞlm formation and bioÞlm maturity, which again is dependent on time of growth, type and frequency of nutrient addition, and the substrate (surface to attach to). Recent studies with young and old bioÞlms grown from oral bacteria have shown that the bioÞlm bacteria were sensitive to NaOCl, chlorhexidine, and iodine for the Þrst 2 weeks of growth [41, 44]. After 3 weeks, the bioÞlms became very resistant to these same agents, in the same concentrations. Furthermore, bioÞlms grown from different sources showed the same pattern of resistance; bioÞlms from six different donors all became resistant to the three disinfecting agents between 2 and 3 weeks of growth [44]. These results make it easy to understand the wide variation of the results in many of the earlier studies with dentin blocks and extracted teeth.

In Vivo Models

Studies done in vivo have the great advantage that real environmental factors are present. These include anatomy, temperature, nutrients, chemistry of the tooth and the periapical area, tissue exudate, host defense, and ÒnaturalÓ bioÞlm. However, many of these factors show great variation from one tooth to another. By selecting only certain teeth, such as the maxillary central incisors, the impact of some factors such as anatomy is reduced. To balance the differences between study groups, a large sample size is usually required, which makes these studies difÞcult to do because of increasing costs and the time required to collect a large enough group of patients. In vivo studies also have ethical limita-

tions which in vitro studies often do not have. In patients, for example, it is not possible to have standardized infections. This could be possible to some extent in animal studies, but animal studies nowadays face strict ethical considerations and high cost. Another important aspect in animal experiments is that the anatomy of the root canal system is different from human teeth [48Ð52].

Although there are many challenges facing in vivo studies on endodontic irrigation and disinfection, this should be the ultimate type of study in the search for optimal treatment protocols. It is clear, however, that when new irrigating solutions or irrigation technologies are introduced, they cannot readily be tested by an extensive in vivo study. Instead, relevant in vitro and ex vivo models with strict control of confounding factors should be used in screening for the best candidates for the in vivo studies.

Sampling Methods

Comparison of the antimicrobial effect of different irrigating solutions and other disinfecting agents has often been done by culturing the bacteria at various stages of the experiment or antimicrobial treatment [53Ð55]. Sampling of the microbes has been done by paper points, endodontic Þles, or by aspirating the sample ßuid from the root canal. The CFU measurement provides information on the amount of viable bacteria one is able to collect in the sample. However, commonly used sampling methods are best suited for planktonic bacteria and bacteria that are only loosely attached to bioÞlm. Sampling with paper points is unlikely to effectively collect bacteria from a bioÞlm. In addition, paper points and Þles only go where Þles used for instrumentation have created the path and space. Untouched areas are likely left untouched by the sample collecting instrument. To increase the possibility of also obtaining some of the ÒhiddenÓ microbes, agitation of sample ßuid by sonic or ultrasonic energy has been used [56Ð58], but their effect on bioÞlm bacteria is questionable. In some in vitro studies, the whole dentin block

has been frozen, pulverized, and cultured in an effort to capture all microbes in the specimen [59, 60].

Results obtained by culturing from direct contact tests using planktonic cultures often show great differences with statistical signiÞcance between different groups [29, 31, 35, 61]. The reason for this may be that the dynamics (speed) of killing planktonic bacteria by different agents typically results in differences in CFUs of even several logarithmic steps [29, 31, 62, 63]. However, culturing from the root canal (in vivo bioÞlms) is a very different situation and is complicated by a number of confounding factors. If the differences in killing root canal bacteria are not great, inherent variations due to the method make it difÞcult or impossible to obtain statistically signiÞcant differences. Recently, confocal laser scanning microscopy together with viability staining has been employed to quantitate the killing of bacteria in the bioÞlm, root canal, and infected dentin [39, 41, 44, 64]. This approach brings promising advantages for the study of the antimicrobial effectiveness of irrigating solutions against microbes in endodontic bioÞlms.

Culturing method only detects those bacteria that are able to grow and form colonies on solid laboratory media and whose growth requirements are supported by the culture medium and growth atmosphere selected. In vitro studies have demonstrated the ability of multiple bacteria to form

a bioÞlm architecture on root canal walls [65Ð 67]. BioÞlm microbes show much greater resistance to antimicrobial agents than planktonic, Òfree-ßoatingÓ microbes [68, 69]. This raises concerns about the validity of laboratory studies based on cultures.

Uninstrumented Parts of the Root

Canal System

The irrigating solutions must be in direct contact with the root canal wall to be effective. This is particularly important in the apical part of narrow root canals. It is well documented that in many teeth 35Ð53 % of the canal wall area, especially in the apical third but also in ribbon-shaped and oval canals, are not touched by the instruments [70Ð74] (Fig. 4.1). Therefore, microbes in these locations have a better chance of surviving. Residual bacteria are commonly found in such hard-to-reach spaces and in lateral canals and dentin canals. In the main root canal, the bioÞlm which is touched by the instruments is likely to be removed, although some of the bacterial cells may become embedded within the smear of tissue [75]. Contrary to this, bioÞlms on the uninstrumented areas remain undisturbed by the mechanical action. The uninstrumented surfaces should therefore always be regarded as

Fig. 4.1 Root canal anatomy of maxillary Þrst molar and the effects of instrumentation as revealed by microcomputed tomography. The preoperative canal system is

shown in red; the post-instrumentation shape of the canal is indicated by green (Courtesy ÒVisual Endodontics/ Artendo Enterprises Inc.Ó)