Determinants of interproximal marginal excess in Class II composites

Increasing use of posterior resin composite has resulted in widespread usage of sectional matrices which aid in proper embrasure formation and adequate contact strength.  Minimal interproximal marginal excess is desirable as considerable time is required for removal during finishing and the interproximal location makes it difficult to detect.  If undetected, or incompletely removed, excess can adversely affect the periodontium. 

Objectives: This study measured the effect of preparation width, clinician's experience level, and matrix system on proximal margin excess (mm²) in Class II posterior composites.

Methods: 180 MO Class II preparations, differing only in width (prepn: 90-narrow, 90-wide), were custom milled in dentoform teeth (KaVo #19).  Matrix systems (matrix: P=Palodent, T=Composi-Tight, V=V-Ring.  Six operators (operator: predoctoral-inexperienced, or postdoctoral-experienced, 3/group) completed 30 restorations (3 matrix x 2 prepn x 5 samples/group) in a randomized order.

Stereomicroscope images (final magnification 80x) of preparations/restorations were analyzed with ImageJ software (NIH) for mm² of marginal excess-flash beyond interproximal margins.  Three-way ANOVA (preparation, matrix, operator) with a Tukey post hoc was used (p< 0.05).

Results: The three-way ANOVA indicated significant main effects for operator (p=0.0031) and matrix (p=0.0055), and a significant interaction between operator and preparation width (p=0.0233).  When analyzed by two-way ANOVA with post hoc test, significant main effects were as follows: matrix within experienced operator group (greater flash with matrix-P); operator within narrow preparations (greater flash with inexperienced group); matrix within wide preparations (greater flash with matrix-P); preparation width and operator within matrix-V with an interaction between width and operator (experienced operators had less flash with narrow preparations, no difference with wider preparations).

Conclusions: Main effects were associated with operator experience and matrix.  Within specific groups inexperienced operators and matrix-P resulted in greater marginal excess while experienced operators and matrix-V resulted in less marginal excess.

T32DEO14678-05