P. E. Murray, A. J. Smith, L. J. Windsor & I. A. Mjor
Oral Biology, Indiana University School of Dentistry, Indianapolis, IN, USA.
Oral Biology, School of Dentistry, The University of Birmingham, Birmingham, UK.
NIOM, Scandinavian Institute of Dental Materials, Oslo, Norway.
College of Dentistry, University of Florida, Gainesville, FL, USA.

The importance of the remaining dentine thickness (RDT) underlying cavity preparations in modifying pulp responses to dental materials has been a topic of controversy for more than a century (Stanley et al. 1975). Over the years, the estimated value of the minimal cavity RDT which does not cause pulp injury has been decreasing. Stanley (1994) suggested that a RDT of  2 mm would protect the pulp from injury caused by most restorative materials and procedures. Subsequently, Pameijer et al. (1991) reported during luting procedures that a RDT of 1mmor more would be sufficient to protect the pulp tissue from the cytotoxic effects of zinc phosphate (ZnP) and resin-modified glass ionomer (RMGI). Nevertheless, it was recently suggested that restoring deeper cavity preparations, carefully cut down to 0.5 mm, with zinc oxide eugenol (ZOE), intermediate restorative material (IRM) and calcium hydroxide (Ca(OH)2)/amalgam, appeared to have little effect on underlying odontoblast numbers for up to 381 days following treatment in patients (Murray et al. 2000b). Moreover, it is unclear if the minimum RDT, at which little or no pulp injury can be observed, varies significantly between different patient variables and restorative materials. An important source of pulp injury is likely to be the increased traumatic operative procedures involved in cutting deeper cavity preparations (Stanley 1961, Darvell 1981). Understanding the interactions between RDT and the degree of pulp injury may be important in understanding why some vital teeth exhibit clinical symptoms leading to the need for endodontic treatment (Zollner & Gaengler 2000). Nevertheless, precise information on the role of the cavity RDT in influencing pulpal injury and dentinal repair responses with common restorative materials is limited (Lee et al. 1992, Murray et al. 2001).
The buffering effect of the cavity RDT to provide pulp protection can be expected to vary not only with dentinal tubule distance from the prepared cavity floor to pulp tissue, but also with dentinal tubule permeability. Tubule permeability is an important factor in allowing the progression of caries, bacterial leakage and chemical irritants towards pulp tissue (Mjor & Ferrari 2002). Consequently, the width of affected tubules and the peritubular secretion of dentine in reducing their width to help mediate pulp protection are important. Peritubular secretion takes place throughout life and is most noticeable in the teeth of older patients, and in those that have experienced wear/erosion (Mjor 2002).
In newly erupted teeth, the most pulpal dentine often has no discernible peritubular dentine lining (Mjor 1966).Therefore, the permeability and peritubular reaction potentials of teeth may be expected to vary with patient age and treatment history, suggesting that although the RDT between patients can be similar, dentine permeability and pulp reactions may be quite different (Mjor et al. 2001). Nevertheless, it is important to characterize human pulp responses to RDT following treatment to gain an improved understanding of pulp- dentine responses to cavity preparation and restoration variables.
Evidence suggests that reductions in the RDT of cavity preparations, increasingly make the pulp susceptible to traumatic injuries caused by cavity preparation and restoration events (Santini & Ivanovic 1996, Murray et al. 2000a).Thus, it is important to avoid needless dentine removal during surgery. Nevertheless, often the RDT of cavity preparations will be determined by the extent of disease progression and treatment regime. Accordingly, it is important to quantify the degree of pulp injury in response to a range of cavity RDTs in combination with common restorative materials, such as Ca(OH)2/amalgamand ZOE. Few studies have attempted to quantify pulp injury as a function of RDT, restoration materials, or other cavity restoration variables, such as cavity width and cavity wall depth. Nevertheless, minimizing pulp injury following treatment and preserving the numbers of pulpal cell populations, particularly the odontoblasts, can be critical to maintain function and vitality of the pulp, as well as reducing the probability of postoperative complications (About et al. 2001a).
The aims of this study were to quantify pulp responses to cavity RDT by measuring and correlating patient variables such as age, cavity preparation variables such as tooth type, tooth surface of cavity preparation, tooth dentine thickness, cavity floor width and cavity wall depth, as well as restoration variables, including the time elapsed since restoration and type of restoration material. Pulpal responses were measured using odontoblast and subodontoblast cell numbers, pulpal inflammation, as well as dentine bridge and reactionary dentine formation.

Patient sample and cavity preparation.
Thirty-one healthy patients, aged between 10 and 16 years of age (mean11.9 years), had 98 class V cavities prepared in teeth, which were scheduled for extraction for orthodontic reasons at the Oslo School Dental Service, Norway. Patient and parental informed consent was provided for the teeth to be used for research purposes. Atotalof49noncarious intact first or second maxillary or mandibular premolars were obtained from the archives of one of the authors (IAM). Cavity preparations were cut into both the buccal, lingual or occlusal dentine using a dental hand-piece with an abundant water spray coolant. Teeth were restored with amalgam, or amalgam lined with Ca(OH)2 or ZOE (Table 1).Teeth were extracted under local anaesthesia after 3-89 days (mean 26 days) observation periods. Some of these teeth have previously been used to study changes in dentine following operative procedures (Mjor 1967a,b). All experiments were undertaken with the understanding and consent of each subject.

Table 1. Numbers of cavity preparations examined according to cavity remaining dentine thickness and restorative material.

Histomorphometric analysis.
Histological axiobuccolingual tooth sections (5 mm) were prepared through experimental areas of all teeth. Histomorphometric analyses were carried out on haematoxylin and eosin-stained sections with an eye piece graticule x50 magnification. Briefly, the cavity RDT was measured as the shortest distance between the mid-point of the cavity preparation floor and the pulp dentine border (Stanley 1968). The cavity wall depth (CWD) was calculated by measuring the distance between the cavity floor and the outermost cut dentine on both sides of the cavity preparation. These measures were divided by two, to provide a single mean of the cavity wall depth per cavity preparation (Murray et al. 2000 a,b,c; 2001, About et al. 2001a).

Reactionary and reparative forms of tertiary dentine secretion.
Reactionary dentine, sometimes called irregular dentine, was identified as an area of increased tertiary dentine secretion with a tubular continuity with the physiological secondary dentine (Mjor1983). The secretion of reparative dentine was formed by a new generation of odontoblast-like cells and did not have a tubular continuity with the physiological secondary dentine (Smith et al.1995). A differentiation between reactionary and reparative dentine was made during the examination of the sections. The dentine at the interface between the primary-secondary dentine continuum and the reparative tertiary dentine has a variable and irregular structure. It is often a tubular as judged by light microscopy and it has been referred to as interface dentine (Mjor 1983). Dentine bridge is a specialized type of reparative dentine secretion because it has a tubular continuity with the newly differentiated odontoblastlike cells, but not the physiological secondary dentine or reactionary and reparative dentine.

Pulpal inflammatory activity.
Pulpal inflammatory activity was categorized according to defined criteria (Mjor & Tronstad 1972, Heyeraas et al. 2001).

• No reaction: The tooth pulp contained no or very few inflammatory cells, subjacent to the cavity dentinal tubules. All cell strata and predentine width appeared uniform and no hyperaemia or haemorrhage was identified. Slight reaction was characterized by an influx of small numbers of cells, including small round lymphocytes, into the cell-free zone in the area adjacent to the cavity dentinal tubules. Some minor irregularities could be observed in the odontoblast layer and predentine thickness.
• Moderate reaction: The odontoblast layer was present, but partly disrupted by hyperaemia, haemorrhage and polymorphonuclear leucocyte, or mononuclear lymphocyte infiltration subjacent to the cavity tubules. The predentine was absent or reduced in width.
• Severe reaction: The odontoblast layer was destroyed by localized abscess formation, and by an intense infiltration of polynuclear leucocyte lesions extending from the cavity tubules. Hyperaemia was found in the tissue surrounding the intense cellular infiltration. The predentine was reduced in width or absent.

Number of odontoblasts.
The numbers of odontoblasts and cells of the subodontoblast layer were counted per mm at the pulp-dentine border beneath the cut dentinal tubules of the cavity preparation. These cell numbers were also counted 0.5 mm away in three independent positions for comparison: immediately occlusal to the cavity tubules, immediately apical to the cavity tubules and also directly opposite the cavity on the adjacent side of the tooth. These cell count locations are shown schematically in Fig.1. The mean from these three areas was calculated to provide a measure of cell numbers independent of the cavity preparation.

Figure 1. Schematic representation of the cell count locations within teeth.

Statistical analysis.
The raw numerical data was examined using one-way analysis of variance (anova) statistics at a confidence level of 95% (StatView software, SAS Inc., Cary, NC, USA).This statistical procedure is amongst the most versatile and conservative of the multiple comparison tests (Dawson-Saunders & Trapp1994).

Remaining dentine thickness of cavity preparations Pulp responses to cavity preparation and restoration are the summation of the interplay between these variables as well as patient factors. However, the effects of cavity RDT on individual factors do not appear uniform because variations in the degree of correlation between RDT and individual factors ranged between P ј 0.0001and 0.9901. RDT was not found to be statistically correlated with most of the variables examined (Table 2).

Cavity wall depth.
The close similarity between the thickness of dentine in premolars and the reproducible positioning of cavity preparations appears to have caused a reciprocal relationship between RDT and CWD (Fig. 2). This means that any proportional increase or decrease in CWD is matched by an opposite equivalent effect on RDT. For example; a lingual CWD of1mm is equivalent to a RDT of 1.04 mm because the addition of both the CWD and RDT always add up to a value of 2.04 mm (_95% confidence interval of 1.817-2.264) (Fig. 2). This suggests the measurement of CWD may provide a high degree of accuracy for estimating cavity RDT, although some variations were observed between occlusal, buccal and lingual cavity preparations (Fig. 2).

Reactionary and reparative dentine secretion.
Dentine bridge formation was highly correlated to RDT (Table 2). This is because the secretion of reparative dentine was observed to take place with a mean pulp exposure depth of _0.15 mm, whilst no reparative dentine was observed in non-pulp-exposed cavities (Fig. 3). This study can only provide very limited information because 10 pulp exposures were examined and only three of these contained any reparative dentine secretion. In each case, the dentine bridge deposition was small. Reactionary dentine secretion appeared to be a more generalized non-pulp-exposed response with a mean RDT of 0.77 mm (Fig. 3); the secretion of reactionary dentine did not appear to be significantly correlated with RDT (Table 2). Small quantities of reactionary dentine were observed in seven from 88 non-pulp-exposed cavities.

Number of odontoblasts and subodontoblast cells.
The number of odontoblasts beneath cavity preparations was highly correlated to the RDT (Table 2). With a decreasing RDT, underlying odontoblast numbers decreased in an exponential manner rather than a linear manner. Compared with unaffected odontoblasts, the numbers were reduced by approximately13.6%beneath a RDT of 2.5-0.5 mm, 33.7% beneath a RDT of 0.5- 0.01mm and 99.0% beneath pulp exposures (Fig. 4). The subjacent subodontoblast cells were less responsive to reductions in RDT (Table 2); only following pulp exposure, did the subodontoblast numbers decrease by _38.4% (Fig. 4).

Restorative materials and reactionary dentine secretion.
The restorative materials were not found to be different from each other in respect to cavity RDT (Table 2); however, the combination of RDT and restorative material did appear to have some effect on reactionary dentine secretion. Cavity restoration with amalgam alone was not associated with reactionary dentine secretion (Fig. 5). ZOE was increasingly associated with reactionary dentine secretion when the RDT was reduced (Fig. 5). The RDT did not appear to influence reactionary dentine secretion following restoration with calcium hydroxide (Fig. 5). Information is limited about the effects of reactionary dentine because only seven cases were observed from 88 cavities, and the quantity of reactionary dentine was small.

Pulpal inflammation.
Little difference was found between the RDTs of none, slight and moderate categories of inflammation (Table 2) (Fig. 6).The severe category of inflammation is of limited usefulness in this study because only one pulp was assigned to this category.

Table 2. Variables correlated to the remaining dentine thickness of cavity preparations.



Figure 2. Regression analysis of cavity wall depth and remaining dentine thickness.



Figure 3. Dentine bridge formation and cavity remaining dentine thickness.



Figure 4. Odontoblast and subodontoblast cell numbers with cavity remaining dentine thickness.



Figure 5. Restorative materials and presence or absence of reactionary dentine secretion with cavity remaining dentine thickness.



Figure 6. Pulp inflammation and cavity remaining dentine thickness.

Discussion.
Radiographic plates and endocator equipment can be used to provide estimates of cavity RDT, but these measurements do not reflect the physiology of dentine. Measurement of dentine tubules along their length appears to represent the most logical and physiologically relevant method (Langeland & Langeland 1966). In this study, RDT was measured as the shortest distance from the cavity floor to the pulp tissue, primarily because of ease of use and avoidance of measuring complications between different points. The advantage of measuring RDT using this method is that it can be used in conjunction with conventional clinical techniques. Although a relationship probably exists between RDT and tubule length, especially in deep cavities, there is a clear need to evaluate pulp responses in conjunction with dentinal tubule length in future studies. Nevertheless, this study has shown how the CWD can be used to estimate RDT because of the uniformity of premolar dentine thickness together with careful cavity positioning. In lingual, occlusal and buccal cavities, the maximum dentine depth of cavity preparation before pulp exposure takes place was found to be 2.33, 1.96 and1.1mm, respectively. Whatever approach is used to estimate the RDT of cavity preparations prior to their restoration, it is clear that the interactions between RDT, cavity preparation and restoration variables are important to the understanding of the injuries sustained to the dental pulp, and its capacity to mediate dentine repair and remain vital throughout life. The basis for a biological approach to clinical practice would be to identify and attempt to reduce sources of tissue injury, such as using the pulp-protective effect of dentine beneath the cavity floor as well as incorporating the dentine repair capacity of teeth as part of treatment. Several reaction patterns of dentine repair have been demonstrated that may be put into clinical use directly (Mjor 1985, 2002). This requires an understanding of the pulpal responses to varying cavity RDTs and to the reactivity and reaction patterns of primary dentine. Studies of human teeth have shown a high prevalence of histological responses to cavity preparation and restoration events. Dentine bridge formation has been observed in 90% of teeth following pulp exposure (Baume 1980) and reactionary dentine secretion has been observed beneath more than 50% of restorations (Murray et al.2000a,b). Severe categories of pulp inflammation have been associated with low RDTs (Hebling et al.1999). In this present study, only three of10 exposed pulps were associated with dentine bridge secretion and only seven of 88 (8%) non-pulp-exposed cavity preparations were associated with reactionary dentine secretion. A possible explanation for this variation may be due to differences in the observation periods. The mean observation time was approximately 27 days in this study versus 46 and 64 days in other studies similarly evaluated (Murray et al.2000a,b).The shorter time frame of this present study would have provided less time for the dentine repair processes to be fully accomplished. Other explanations for these differences include a variation between patient samples, including the possible effects of treatment history and other factors such as age. The prevalence and magnitude of pulpal responses could theoretically be dependent on subjective criteria that are difficult to measure quantitatively, such as variations in the healing response between individual patients and differences in clinical technique between clinicians. Little is known about the importance of these factors and it was not possible to address the mint his present study. Consequently, we have focused on pulp-dentine responses to RDT or pulp exposure, which is somewhat under the control of the clinician.
In an attempt to overcome difficulties with the use of subjective qualitative criteria, quantitative approaches were used, whenever possible, to assess pulp activity under controlled conditions of patient age, health and specimen tissue preparation. This is because qualitative data always contains an element of subjectivity (Warfvinge1987). The application of the anova multiple comparison test avoided the problem of having to compute many individual comparison tests between histological raw data (Qvist & Stotlze1982). If this test were not performed, the multiple tests between different pairs of means would alter the a-level, not for each comparison, but for the experiment as a whole. For example, if the six pairs of means shown in Fig. 4 are compared using individual tests, and if each comparison is made by using a ј 0.05, there is a 5%chance that each comparison will falsely be called significant, i.e. a type I error would occur six different times. Overall, therefore, there is a 30% chance (6 _5%) of declaring one of the comparisons incorrectly significant. The use of nova by the authors overcame these problems and in agreement with Dawson- Saunders & Trapp (1994) provided a versatile and conservative statistical method.
Many factors have been suggested to injure the pulp, including cavity preparation trauma, restorative dental procedures, operator hand instrumentation, systemic diseases, caries, attrition, erosion, chemicals, dental materials and bacterial leakage (Diamond et al. 1966, Baume 1980, Cox et al. 1987; 1992, Magloire et al. 1992, Lesot et al.1994, Smith et al.1994;1995). Often, the factors correlated with pulp injury have proved to be controversial. Clearly, the RDT or proximity of the injury stimulus to the pulp tissue will have some impact on the severity of pulp response. Often a reduction in proximity of the cavity floor to the pulp tissue has been reported to increase reactionary dentine deposition (Murray et al. 2000b), whilst other studies have reported little correlation between RDT and reactionary dentine activity (Stanley 1961, Santini & Ivanovic 1996). The present study did not demonstrate a relationship between the RDT and the presence of reactionary dentine. This lack of correlation may be related to the minimal nature of the responses observed, resulting from the short postoperative extraction schedule. However, the formation of a dentine bridge was correlated to the creation of a pulp exposure. This would suggest an accord with previous investigations that cavities prepared close to the pulp tissue can injure the underlying odontoblast cells (Darvell1981, Lee et al.1992). Injury to the odontoblasts may explain the loss of capacity of these cells to secrete reactionary dentine, and the secretion of dentine bridge by a new generation of odontoblast-like cells (Smith et al.1994).
The effect of restorative materials on pulpal activity has similarly proved to be controversial. Some types of restorative materials, such as Ca(OH)2-based products have been reported to stimulate pulpal repair activity by increasing reactionary dentine deposition (Stanley 1968), whereas some other studies have shown the effects of the restorative material on pulp responses are minimal (Cox 1992). This present study did not demonstrate a relationship between reactionary dentine secretion and the pure form of Ca(OH)2 mixed with water as a material liner. This may be due to the precipitation of crystalline salts within the dentinal tubules, reducing their permeability and preventing any adverse reactions in the pulp (Mjor & Ferrari 2002).The results from this study were also in agreement with previous observations, indicating the negative effect of ZOE (Stanley 1968, Kirk & Meyer 1992) and amalgam on reactionary dentine secretion. A smaller RDT was required to observe reactionary dentine with ZOE and no reactionary dentine was associated with amalgam, although amalgam was only used to restore very shallow cavity preparations. Therefore, the larger RDT may be responsible for the lack of reactionary dentine, rather than the chemical activity of amalgam, because reactionary dentine activity is reduced with increasing RDT (Stanley et al. 1966). The therapeutic significance of why some restorative materials may be associated with larger areas of reactionary dentine is difficult to explain because of the limited information available.
However, changes in the buffering effect of dentine with different RDTs can influence the cytotoxicity and chemical activity of restorative materials on underlying pulp tissue (Murray et al. 2000c). Alterations in dentine physiology and chemistry in differing tooth locations may also be able to explain some differences. For example, Ca(OH)2 on primary dentine will quickly cause precipitation of mineral salts within the tubules (Mjor et al. 1961, Mjor & Furseth1968) and that will reduce dentine permeability and influence pulp tissue reactions. This biologically positive effect can be used to prevent reactions from restorative materials when Ca(OH)2 is used as a liner (Mjor1963).
Some areas of coronal dentine show marked variation in the structure of the tubules, e.g. the tubules in the most pulpal dentine in newly erupted teeth have no discernible peritubular dentine lining them whilst those in the bulk of dentine are relatively uniform and are lined by peritubular dentine (Mjor 1966). The structure and reaction potentials of this dentine will vary depending on the age of the patient and wear/attrition amongst other factors that may have affected the tissue. In teeth from young individuals, the dentine will be truly unaffected and therefore more permeable than the dentine in the central part of the cavity preparation, which may have mineralized due to caries. This situation calls for particular attention to the protection of the most peripheral part of cavity preparations, and of large parts of crown preparations, especially those on intact teeth of young individuals (BjMrndal & Mjor 2001).
Cavity positioning is also important because the area of tubules exposed by the preparation can vary from about 80%of the total area in deep preparations in teeth from young individuals to 4% in preparations in the peripheral part of the dentine in teeth from individuals in any age group (Ketterl1961). Consequently, the buffering effect of dentine tubule structure in preventing the progression of caries, leakage of bacteria and penetration of chemicals, appears to be just as important as the RDT in providing pulp protection (Mjor & Ferrari 2002).
The combined effects of bacteria and their toxins has been strongly associated with increased pulpal inflammatory activity. This inflammatory activity has been associated with injury to pulpal tissue and reductions in the numbers of vital cell populations (Brannstrom 1984, About et al. 2001b). The absence of bacteria at the cavity margins in this present study probably explains the lack of pulpal inflammatory reactions and the failure to identify any significant correlation between cavity RDT and pulpal inflammation. This suggests that pulpal inflammation is not primarily stimulated by the effects of cavity preparation followed by restoration with Ca(OH)2, amalgam or ZOE. However, the RDT of preparations may be critical in modifying pulpal responses to more injurious restorative materials, such as adhesives (Hebling et al. 1999) and also in the presence of caries and bacteria.
This study has provided evidence to highlight the complex interplay between restorative materials, cavity RDT and cavity dimensions which can interact together to injure the pulp tissue. Clearly, cavity RDT plays a central role in determining the extent of pulpal injury and repair responses. The mechanisms involved in the detection of dentine damage, its transduction to the odontoblasts and the stimulation of increased odontoblast dentine synthesis and secretion remain incompletely understood (Kardos et al. 1998). However, damage to the odontoblastic process, injury to the intratubular elements, nerve damage and the presence of bioactive molecules solubilized from the injured dentine matrix, may be involved (Smithet al.2001).Currently, following cavity restoration, a high proportion of teeth with vital pulps exhibit symptoms requiring endodontic treatment (Zollner & Gaengler 2000). Although the mechanisms of odontoblast peritubular and tertiary dentine secretion remain incompletely understood, the ability to anticipate and manipulate its activity should form part of treatment planning. The basis for a biological approach to clinical practice would be to identify and attempt to reduce sources of tissue injury, and also exploit the dentine repair capacity of teeth intentionally as part of treatment. This should help to minimize postoperative complications following restorative treatment. An improved understanding of the degree of pulp responses to varying RDTs will help with the identification of pulpal complications.

References.

About I, Murray PE, FranquinJ-C, Remusat R, Smith AJ (2001a)   The effect of cavity restoration variables on odontoblast cell numbers and   dental repair. Journal of Dentistry 29, 107-17.
About I, Murray PE, FranquinJ-C, Remusat M, Smith AJ (2001b) Pulpal inflammatory   responses following noncarious class Vrestorations. Operative Dentistry   26,336-42.
Baume LJ (1980) The Biology of Pulp and Dentine. Monographs in Oral Science.   Basel, Switzerland: Karger,159-82.
BjMrndal L, Mjor IA (2001) Pulp-dentin biology in restorative dentistry.   Part 4. Dental caries - characteristics of lesions and pulpal reactions.   Quintessence International 32, 717-36.
Brannstrom M (1984) Communication between the oral cavity and the dental   pulp associated with restorative treatment. Operative Dentistry 9,57-68.
Cox CF, Keall CL, Keall HJ, Ostro E, Bergenholtz G (1987) Biocompatibility   of surface-sealed dental materials against exposed pulps. Journal of Prosthetic   Dentistry 57,1-8.
Cox CF (1992) Microleakage related to restorative procedures. Proceedings   of the Finnish Dental Society 88,83-93.
Cox CF, White KC, Ramus DL, Farmer JB, Snuggs HM (1992) Reparative dentine:   factors affecting its deposition. Quintessence International 23, 257-70.
Darvell BW(1981) Effect of dentine thickness on pulpal changes beneath restorative   materials. Australian Journal of Dentistry 26,80-1.
Dawson-Saunders B, Trapp RG (1994) Basic and Clinical Biostatistics,   2nd rev. edn. Connecticut: Appleton & Lange, 125-43.
Diamond R, Stanley HR, Swerdlow H (1966) Reparative dentine formation resulting   from cavity preparation. Journal of Prosthetic Dentistry16,1127-34.
Hebling J, Giro EMA, Costa CAS (1999) Human pulp response after an adhesive   system application in deep cavities. Journal of Dentistry 27,557-64.
Heyeraas KJ, Sveen OB, Mjor IA (2001) Pulp-dentine biology in restorative   dentistry. Part 3. Pulpal inflammation and its sequela. Quintessence International   32,611-25.
Kardos TB, Hunter AR, Hanlin SM et al. (1998) Odontoblast differentiation:   a response to environmental calcium? Endodontics and DentalTraumatology14,105-11.
KetterlW(1961) Studie uber das dentin der permanenten zahne des menschen.   Stoma,14,79-96.
Kirk EEJ, MeyerMJ (1992) Morphology of the mineralizing front and observations   of reparative dentine following induction and inhibition of dentinogenesis in   the rat incisor. Endodontics and Dental Traumatology 8,195-201.
Langeland K, Langeland LK (1966) Pulp reactions to crown preparation, impression,   temporary crown fixation, and permanent cementation. Journal of Prosthetic   Dentistry 15, 129-143.
Lee SJ, Walton RE, Osborne JW (1992) Pulp response to bases and cavity depths.   American Journal of Dentistry 5,64-8.
Lesot H, Smith AJ, Tziafas DB, Begue-Kirn C, Cassidy N, Ruch JV (1994) Biologically   active molecules and dental tissue repair: a comparative view of reactionary   and reparative dentinogenesis with the induction of odontoblast differentiation   in vitro. Cells and Materials 4,199-218.
Magloire H, Bouvier M, Joffe A (1992) Odontoblast response under carious   lesions. Proceedings of the Finnish Dental Society 88, 257-74.
Mjor IA (1963) The effect of calcium hydroxide, zinc oxide/ eugenol and amalgam   on the pulp. Odontol Tidskr 71, 94-105.
Mjor IA (1966) Microradiography of human coronal dentine. Archives in   Oral Biology11, 225-34.
Mjor IA (1967a) Histologic studies of human coronal dentine following the   insertion of various filling materials in experimentally prepared cavities.   Archives in Oral Biology 12, 441-52.
Mjor IA (1967b) Histologic studies of human coronal dentine following cavity   preparations and exposure of ground facets in vivo. Archives in Oral Biology12,   247-63.
Mjor IA (1983) Dentin and pulp. In: Mjor IA, ed. Reaction Patterns in   Human Teeth. Boca Raton, FL: CRC Press,63-156.
Mjor IA (1985) Dentin-predentine complex and its permeability: pathology   and treatment overview. Journal of Dental Research 64 (Special Issue),621-7.
Mjor IA (2002) Clinical management and tissue changes associated with wear   and trauma. In: Mjor IA, ed. Pulp-dentine biology in restorative dentistry.   Carol Stream, IL: Quintessence Publishing,77-94.
Mjor IA, Ferrari M (2002) Pulp-dentine biology in restorative dentistry.   Part 6. Reactions to restorative materials, tooth-restoration interfaces, and   adhesive techniques. Quintessence International 33, 35-63.
Mjor IA, Finn SB, Quigley MB (1961) The effect of calciumhydroxide and amalgam   on noncarious, vital dentine. Archives in Oral Biology 3, 283-91.
Mjor IA, FursethR(1968) The inorganic phase of calciumhydroxide and corticosteroid   covered dentine studied by electron microscopy. Archives in Oral Biology13,755-63.
Mjor IA, Sveen OB, Heyeraas KJ (2001) Pulp-dentin biology in restorative   dentistry. Part1.Normal structure and physiology. Quintessence International   32, 427-46.
Mjor IA, Tronstad L (1972) Experimentally induced pulpitis. Oral Surgery   34,102-8.
Murray PE, About I, Franquin J-C, Remusat M, Smith AJ (2001) Restorative   pulpal and repair responses. Journal of the American Dental Association132,   482-91.
Murray PE, About I, Lumley PJ, FranquinJC, Remusat M, Smith AJ (2000a) Human   odontoblast cell numbers after dental injury. Journal of Dentistry 28, 277-85.
Murray PE, About I, Lumley PJ, Smith G, FranquinJ-C, Smith AJ (2000b) Postoperative   pulpal and repair responses. Journal of the American Dental Association   131,321-9.
Murray PE, Lumley PJ, Ross HF, Smith AJ (2000c) Tooth slice organ culture   for cytotoxicity assessment of dental materials. Biomaterials 21,1711-21.
Pameijer CH, Stanley HR, Ecker G (1991) Biocompatibility of a glass ionomer   luting agent. 2. Crown cementation. American Journal of Dentistry 4,134-41.
QvistV, Stotlze K(1982) Identification of significant variables for pulp   reactions to dental materials. Journal of Dental Research 61, 20-4.
Santini A, IvanovicV (1996) The quantification of tertiary dentine formation   in response to materials commonly placed in deep cavities in general practice   in the UK. Primary Dental Care 3,14-22.
Smith AJ, Cassidy N, Perry HB, Begue-Kirn C, Ruch J-V, Lesot H (1995) Reactionary   dentinogenesis. International Journal of Developmental Biology 39, 273-80.
Smith AJ, Cassidy N, Plant CG et al. (1994) Odontoblast stimulation in ferrets   by dentine matrix components. Archives in Oral Biology 39,13-22.
Smith AJ, Tobias RS, Murray PE (2001) Transdentinal stimulation of reactionary   dentinogenesis in ferrets by dentine matrix components. Journal of Dentistry   29,341-6.
Stanley HR (1961) Traumatic capacity of high-speed and ultrasonic dental   instrumentation. Journal of the American Dental Association 63,750-66.
Stanley HR (1968) Design for a human pulp study. Part 1. Oral Surgery,   Oral Medicine and Oral Pathology 25,633-47.
Stanley HR, White CL, McCray L (1966) The rate of tertiary (reparative) dentine   formation in the human tooth. Oral Surgery, Oral Medicine and Oral Pathology   21,180-9.
Stanley HR, Conti AJ, Graham C (1975) Conservation of human research teeth   by controlling cavity depth. Oral Surgery, Oral Medicine, Oral Pathology   39,151-6.
Stanley HR (1994) Dental iatrogenesis. International Dental Journal 44,3-18.
Warfvinge J (1987) Morphometric analysis of teeth with inflamed pulp.   Journal of Dental Research 66,78-83.
Zollner A, Gaengler P (2000) Pulp reactions to different preparation techniques   on teeth exhibiting periodontal disease. Journal of Oral Rehabilitation   27,93-102.