Predictability of anthropomorphic measurements in implant selection for breast reconstruction: a retrospective cohort study.

[BACKGROUND] Preoperative implant planning for breast reconstruction is often at risk of being changed perioperatively. This study examined which factors are associated with a change of implant selection.

[METHODS] Women who had unilateral two-stage breast reconstruction between 2002 and 2007 were studied. Inclusion criteria were photographic evidence of preoperative skin markings indicating breast dimensions and a selected implant model. Multivariable logistic regression was used to identify variables associated with a changed selection.

[RESULTS] Among the 496 women studied, 308 preoperative implant choices (62.1%) were changed during surgery. A change in plan was significantly associated with symmetrization surgery involving contralateral reduction mammaplasty (OR = 1.92; 95% CI, 1.12 to 3.29) and contralateral mastopexy (OR = 2.26; 95% CI, 1.29 to 3.96), but not with BMI. The required implant width changed more than 0.5 cm in 70 cases (14.1%) while height changed more than 0.5 cm in 215 cases (43.2%). The likelihood of a change was high for large preoperative widths (OR = 9.66 for 15.5 cm) and small preoperative heights (OR = 2.97 for 10.5 cm). At a mean follow-up of 16.6 months, patient satisfaction was good or average in 92.1% of cases and 5.9% of implants had been replaced with another model, indicating that the perioperative implant selection was usually appropriate.

[CONCLUSIONS] This study documents the frequency with which implant choices, despite accurate preoperative planning, are changed perioperatively as a result of relatively small differences in anthropomorphic measurements. Perioperative recalculation of breast dimensions may have an advantage in terms of patient reoperation rates. Changes in width were less frequent than changes in height and projection. Contralateral surgery, large width, and small height were the most influential factors. Level of Evidence: Level IV, risk / prognostic study.