110 Chapter 7 as measured by the frequency of return visits for treatment of complications and additional procedures.(25) The Dutch national guideline recommends preferably not to perform immediate implant-based breast reconstruction (DTI) in patients with more than two of the following preoperative risk factors: smoking, BMI ≥30 kg/m2, bilateral surgery, age >55 years and larger breasts.(26) This recommendation is, among others, partially based on a large study which performed a risk analysis in 14.585 patients who underwent an immediate breast reconstruction.(3) In this study the overall implant loss rate was very low (0.8%) compared to 6.7% in current study, which can be attributed to the short follow-up time of 30 days as our median time to surgical removal of the implant was 42 days (IQR: 21 to 83 days). As mentioned in the results, a precise risk prediction of implant loss could be calculated for each individual patient, using the given formula. To facilitate an easier risk calculation, BMI was dichotomized to calculate the predicted probabilities of implant loss for each number of risk factors. During the preoperative consultation, the treating physician can clearly outline the estimated risk of implant loss by using these prediction tools, enabling the patient to make a well-informed decision. Other alternatives, such as an autologous reconstruction, a delayed reconstruction, or no reconstruction at all, should be taken into consideration for individuals with an unacceptable high anticipated risk. This nationwide population-based study has several limitations. First of all, the data from the database are all anonymized due to privacy regulations. For this reason, the relatively high percentage excluded reconstructions, because of missing data on one or multiple risk factors could not be decreased. Furthermore, in this cohort, no subjects with four risk factors present were included, therefore, for this number of risk factors, the observed probability could not be calculated. Next, there was a restriction to the data collected from the DBIR database. Since implant loss was not an existing variable, the outcome was composed of multiple variables. Also nipple sparing was not a direct variable and had to be derived from the ‘incision site’. Finally, the DBIR is filled in directly after surgery, therefore the accuracy of adjuvant radiotherapy cannot be guaranteed as the indication for adjuvant therapy may change after postoperative pathology reports or multidisciplinary tumor board meetings. In conclusion, these nationwide population-based DBIR data were applied to create an easy to use risk prediction model for implant loss after immediate implant-based breast reconstruction. Four risk factors were included: BMI, active smoking status, previous radiotherapy and prepectoral placement. The model reported the predicted risk for implant loss for each number of risk factors,
RkJQdWJsaXNoZXIy MTk4NDMw