The distance from the implant-abutment connection to the peri-implant marginal bone level was measured mesially and distally in parallel with the long axis of the implant using open-source software (ImageJ, National Institutes of Health, Bethesda, MD, USA). The distance from the cemento-enamel junction to the marginal bone level at the neighbouring tooth surfaces was also measured in parallel with the root surface [35]. The marginal bone level was defined as the most coronal level of the alveolar bone with a normal width of the periodontal ligament [36]. The correction of magnification was based on the known distance between the implant threads (0.6 mm) or implant length.

All patients were asked to fill out a questionnaire regarding their overall satisfaction with the implant treatment at the time of placement of the implant crown (baseline) and at the final follow-up. Their answers were registered using a 10-cm-long visual analogue scale (VAS) ranging from 0 (indicating discontent with the implant treatment) to 10 (indicating satisfaction with the implant treatment).

Each patient’s record was thoroughly reviewed, and all technical and biological complications during the follow-up period were registered. Two examiners (JH and FI) made all registrations and measurements.

Data analysis

Data management and analysis including calculation of descriptive statistics were performed using Excel (Microsoft, Redmond, WA, USA) and STATA (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC, USA). No power calculation of sample size was included due to lack of relevant data on dental implants and platelet-rich fibrin in previously published studies. Data were analysed using a mixed model for repeated measurements. Comparisons within and between the groups were performed as post-hoc tests following the model. Normality of the residuals (the difference between the actual value of the outcome and the fitted values) and the homogeneity of the variance of the residuals were evaluated using the visual inspection of the QQ-plot of the residuals and a scatter plot of the residuals and the fitted values. The outcome of BOP and REC of the implant were dichotomised into absence or presence and analysed using a generalised linear model with log-link function analysing the ratio of the chance of BOP or REC (generally known as risk ratio). The remaining clinical parameters were tested using a simple linear regression model.

For interobserver repeatability, two observers (JH and FI) analysed the intraoral radiographs of five patients (20 radiographs). Additionally, for assessment of intrapersonal reproducibility, the images of all patients (104 radiographs) were measured twice (JH) allowing for a 3-month interval between the two measurements. The repeatability and reproducibility were described by the intraclass correlation coefficient (ICC) by a two-way mixed-effects model.

A statistically significant difference was considered when p < 0.05.