During preparation of the 96 canals, a total of11 instruments were separated. Therefore, the following results are based on the remaining 85 canals. Five canals with 288 and six canals with 358 curves were excluded.

Instrument failure.
Table 1 details the number of deformations and fractures of instruments that occurred during the study. All failures occurred at the tip region of the instruments. Independent of the curvature of the canals, none of the stainless steel K-Flexofiles separated. In the 288-curved canals, five K3 instruments separated, and in the 358- curved canals six nickel-titanium instruments separated (Table 2). The number of separated instruments (w2 =5.442; P < 0.05) was significantly higher for the K3 files whilst the number of permanently deformed instruments (w2 =0.011; P = 0.918) was not significantly different between the instrument types.

Preparation time.
The mean time taken to prepare the canals with the different instruments is shown in Table 3. Independent of the curvature of the canals, the shortest mean preparation time was recorded when K3 instruments were used. Bothin288-and358-curvedcanals, K3 was significantly faster than K-Flexofiles (P < 0.001).

Change of working length.
All canals remained patent following instrumentation; thus, none of the canals became blocked with resin shavings. None of the canals showed overextension of preparation, whereas a loss of working distance was found in several canals.
The mean loss of working length that occurred with the different instruments is listed in Table 4. The differences between the two instrument types were not statistically significant, in either the 288- (P = 0.074) or 358-curved canals (P = 0.096).

Table 1. Number of fractured and permanently deformed instruments.



Table 2. Number of separated K3 files with respect to size and taper.

Canal shapes.
The results concerning the assessment of canal aberrations are summarized in Table 5.With respect to the different types of aberrations evaluated, both in canals with 288 and with 358 curves, there were no significant differences between the two instrument types (w2-test; P > 0.05), but more zips and more ledges were created with K-Flexofiles.
On average, in the canals with 288 curves, K3 instruments removed material more evenly on the outer as well as on the inner side of curvature (Fig. 3a). With the exception of two measuring points, significant differences (P < 0.05) occurred between resin removal achieved with the two different instruments (Table 6). The canals prepared with K3 instruments remained better centred compared with those enlarged with K-Flexofiles (Fig. 3).
In the canals with 358 curves, K3 instruments removed more material on the outer side of the curvature (Fig. 4a). In average, only limited material removal occurred in the inner side of the curvature in the apical part of the canals (Table 7). Canals shaped with K-Flexofiles had material removed mainly in the last 1-5 mm along the outer side of the curvature, resulting in severe outer widening of the canal (Fig. 4b).This effect was significantly different compared with the material removal achieved with K3 files at these five measuring points (P < 0.001). In general, the K3 instruments showed a more centred enlargement compared with the K-Flexofiles (Fig. 4).

Table 3. Mean preparation time (min) and SD with the two different instruments.



Table 4. Mean loss of working length (mm) and SD with the two different instruments.



Table 5. Incidence of canal aberrations by instrument types.



Figure 3. Mean changes in the canal shape of 288-curved canals as the result of instrumentation with
(a) rotary nickel-titanium K3 instruments and
(b) stainless steel hand K-Flexofiles (n = 24 canals in each case).



Table 6. Mean material removed (mm) and SD at the different measuring points after instrumentation of simulated 288-curved canals.



Figure 4. Mean changes in the canal shape of 358-curved canals as the result of instrumentation with
(a) rotary nickel-titanium K3 instruments and
(b) stainless steel hand K-Flexofiles (n = 24 canals in each case).



Table 7. Mean material removed (mm) and SD at the different measuring points after instrumentation of simulated 358-curved canals.