Results.

**Prepreparation findings and values.**

The autoclave sterilization of the specimens proved to be 100% effective as subsequent checks made for evidence of microbial growth were negative. The average concentration of the inoculate used for reinfecting the specimens was 8.24 ïƒ—ï€ 1012 CFUmLâ€“1 (range 4.27 ïƒ—ï€ 1012â€“1.47 ïƒ—ï€ 1013 CFUmLâ€“1, respectively. The SEM examination revealed that E. faecalis occupied the main root canals in the randomly selected specimens and had penetrated the dentinal tubules to varying extents. Figure 1 is an example of the penetration of dentinal tubules by the bacteria in the apical one-third of a root canal.

Since no other morphological types of bacteria were evident, the results from the SEM suggested that the reinfection procedure was successful.

The results of the E. faecalis sensitivity tests showed that only NaOCl and EDTA exhibited antimicrobial activities as expected. Neutralization of the NaOCl by Na2S2O3 was complete, as the corresponding plate showed no signs of inhibited microbial growth.

**Figure 1.** Representative SEM appearance of root canal wall in the apical one-third after reinfecting with E. faecalis showing the bacteria occupying dentinal tubules (original magnification 6400).

**Prepreparation results.**

There were no statistically significant differences amongst the groups for the prepreparation concentration of microorganisms or working lengths prior to treatment (P = 0.865 and 0.961, respectively) (Table 1).

**Table 1.** Means (standard deviations) of prepreparation concentration and working length by group.

**Table 2.** Specimen number, specimen group number, group, pre- and postpreparation concentrations, percentage reduction, presence (1) or absence (2) of E. faecalis after preparation and 24 h postpreparation neat appearance.

**Postpreparation results.**

Table 2 shows the postpreparation results obtained for each specimen from the three study groups showing the concentrations of E. faecalis, as a percentage of bacterial reduction (calculated using the pre- and postpreparation concentrations), the presence or absence of bacteria after preparation (another interpretation of the latter) and the results of the 24 h appearance.

In group A, 15 specimens (i.e. all but one specimen) were rendered bacteria free after preparation. In Table 2, column 6, the results from group A show 15 specimens with 100% bacterial reduction and one specimen with 99.99% bacterial reduction.

In group B, 13 specimens were rendered bacteria free after preparation with less than 100% reduction in three specimens.

In group C, the positive control group, all six specimens showed less than 100% reduction in E. faecalis after irrigation.

Ten of the neat samples appeared turbid after 24 h incubation, indicating the presence of E. faecalis. The 10 neat samples were from the same specimens (from the three groups) that gave positive results for the presence of E. faecalis after they were prepared (i.e. one specimen from group A, three from group B and six from group C).

**Statistical analysis of the results.**

Table 3 shows the mean percentage reduction for the three groups. All three groups showed high percentage reductions, with group C showing a slightly lower value. There were two other variables that may have also been linked to the percentage reduction in concentration, namely the working length and the prepreparation concentration. A generalized linear model (used for continuous data), was used to identify which, if any, of the variables might have had a significant effect on the percentage reduction in concentration. The prepreparation concentration and the working length were not significant (P = 0.656 and P = 0.616, respectively), whilst the group effect (i.e. the preparation regimen applied to each group) was highly significant (P = 0.007). Thus, the only variable that had a significant effect on the percentage reduction in concentration was the preparation method of the group. Table 4 shows the results of the follow-up multiple paired comparisons, using the Bonferroni Correction (Bland & Altman 1995), which was used to identify between which groups the differences exist. The interval estimates confirm that there was a significant difference between groups A and C (group A had on average a significantly greater percentage reduction in concentration by between 0.05% and 0.30% than group C). There was also a significant difference between groups B and C (group B had on average a significantly greater percentage reduction in concentration, by between about 0.04% and 0.31% than group C). The interval estimate however, showed that there was no significant difference in percentage reduction between groups A and B (interval ranged from â€“0.09% to 0.10%). Thus, the GT rotary preparation with antimicrobial irrigation gave a significantly greater mean percentage reduction in E. faecalis concentration in both experimental groups (A and B), regardless of the size to which the apical region of each specimen was prepared, compared to the control group (C). In addition, there was no significant difference in the percentage reduction in concentration of bacteria between the two experimental groups. The data in Table 2 however, show that one sample in group A was not rendered bacteria free compared with three in group B. Thus, another method was used to consider a second â€˜outcome of interestâ€™.

This â€˜outcome of interestâ€™ was considered to be either the presence (1) or absence (2) (Table 2, column 7) of E. faecalis in the sample after preparation. Therefore:

- presence was defined as postprep. concentration > 0 and was arbitrarily designated 1;
- absence was defined as postprep. concentration = 0 and was arbitrarily designated 2.

In this statistical method, the determination to be made was whether the percentage of â€˜absenceâ€™ was the same across the three study groups. The percentages of â€˜absenceâ€™ in the three groups were 93.75% for group A, 81.25% for group B and 0% for group C as shown in Table 5. The percentages clearly suggested that there was a difference between the experimental and the control groups. A stepwise logistic regression (used for binary data) was used to identify whether any of the explanatory variables (i.e. group preparation regimen, prepreparation concentration and working length) had a significant effect on whether E. faecalis was absent after preparation. In this instance, both the prepreparation concentration and working length were again not significant (P = 0.626 and P = 0.529, respectively), whilst the group effect was highly significant (P < 0.001). Therefore the only significant variable was the preparation method. Intergroup analysis of the results indicated that groups A and C were significantly different, as were groups B and C (i.e. the percentage of â€˜absencesâ€™ significantly different in both instances). However, the outcome from groups A and B showed that the preparation effect was not significant (P = 0.276). Therefore, there is insufficient evidence to reject the null hypothesis that the percentage of â€˜absenceâ€™ in the two experimental groups is the same.

**Table 3.** Means (standard deviations) of percentage reductions in concentration by group.

**Table 4.** Bonferroni-adjusted intervals for differences between groups for percentage reduction in concentration.

**Table 5.** Percentage of 'absence' per group (see Table 2, column 7).