The regional arthroplasty registry RIPO provides active surveillance in the Italian region Emilia-Romagna (around 4,500,000 inhabitants), collecting data on hip, knee, and shoulder arthroplasties and revision surgeries since January 2000 [8]. Involving 68 orthopedic facilities in the region, the registry is a cross-checked database with a reported capture rate of 98%, with the 2% of missing data due to the lack of adherence [8]. RIPO collects the forms filled by all the surgeons performing primary arthroplasty or revision surgeries: the clinical conditions of the patients, the devices (batch and code), and the surgical technique (approach and fixation) are reported.

The RIPO registry was inquired about three stems implanted in primary cementless THAs with two versions, ST and DT. The three involved stems, with the ST and DT versions, were Apta-fix/Apta (Adler Ortho, Milan, Italy), Hydra-fix/Hydra (Adler Ortho, Milan, Italy), and Recta-fix/Recta (Adler Ortho, Milan, Italy) (Fig. 1). In the DT versions (Apta, Hydra, Recta), all the three stems shared the same titanium alloy Ti6Al4V and the same modular junction, Modula (Adler Ortho, Milan, Italy), providing 27 version/offset/length combinations by using 15 Ti6Al4V modular necks (Fig. 2).

The six stems involved in the study were manufactured by Adler Ortho (Milan, Italy): Apta (A), Apta-fix (B), Hydra (C), Hydra-fix (D), Recta (E), and Recta-fix (F)

Preoperative (A, C, E) and postoperative (B, D, F) pelvis X-rays of three clinical cases, one for each stem with the DT version: Apta (B), Hydra (D), and Recta (F)

Apta-fix design is an anatomic stem with an extensive hydroxyapatite coating (classification according to Khanuja (6): it is available in eight sizes, with standard and offset configurations (lateralization of 7.5 mm) [9, 10]. The caput-collum-diaphyseal (CCD) angle changes from the standard configuration (135°) to the offset one (130°) [9]. Both the standard and offset solutions have 6° of neck anteversion [9]. Hydra-fix design is a hydroxyapatite-coated single wedge tapered stem (classification according to Khanuja: 1): it is available in 11 sizes, with standard and offset configurations (lateralization of 7.5 mm) [10, 11]. Standard and offset configurations share the same CCD angle (135°) [11]. Recta-fix is a corundum blasted stem with a tapered rectangular design (classification according to Khanuja: 3C): nine sizes are available, with standard (CCD 131°) and offset configurations (CCD 123°, with 7.5 mm lateralization) [10, 12].

The inclusion criteria were: residing patients (to minimize the loss of patients at follow-up), THAs performed for primary osteoarthritis, use of Delta ceramic-on-Delta ceramic bearings (Ceramtec, Plochingen, Germany), THAs with 3D-printed Fixa TiPor (Adler) socket, modern version of titanium-on-titanium modular junctions (identified by “046XXXX” code).

All the THAs performed for reasons others than primary osteoarthritis, in nonresiding patients, with bearings others than Delta-on-Delta and cups others than TiPor Adler and involving previous versions of the modular junction were excluded.

Every selected design was stratified into two cohorts according to modularity, ST and DT: three pairs of cohorts were eventually identified. Demographics and implant-related features of ST and DT cohorts were collected and compared for every single design. The survival rates of ST and DT cohorts were calculated and compared, using different endpoints. Similarly, adjusted HRs for different reasons for failure were calculated and compared.

Institutional review board approval was waived due to the registry nature of the study and data anonymization.

Statistical analyses were performed using SPSS software (version 14.0.1, Chicago, IL) JMP, version 12.0.1 (SAS Institute Inc, Cary, NC, 1989–2007. Data were provided as raw data, ranges, frequencies, and percentages. Continuous variables of demographic and implant-related features were analyzed using Student’s t-test, and frequencies and percentages using chi-square test. The survival curves were calculated and plotted using the Kaplan–Meier method (time in years on the x-axis and percentage of survived implants on the y-axis): the curve was flanked by a pair of 95% confidence interval curves. The implants were considered “surviving” at the last date of observation (date of death or 3 December 31, 2020) when no single component was replaced. The log-rank test was adopted to test the survival curves (threshold, p = 0.05). A multivariate Cox regression model was used to detect failures, with Wald test to detect any significance. HRs and 95% confidence intervals were specified. Threshold for significance was p = 0.05.

A total of 5789 THAs were included. By stratifying the implants per version, 1984 (75.9%) Apta and 629 (24.1%) Apta-fix; 1672 (70.8%) Hydra and 690 (29.2%) Hydra-fix; 662 (81.3%) Recta and 152 (18.7%) Recta-fix were enrolled. Apta, Hydra, and Recta (the three DT versions) had a female prevalence (respectively 64.5%, 66%, and 59.2%; p < 0.001). The mean age, mean height, and mean weight of the Apta/Apta-fix cohorts were, respectively, 69.6 years (range: 37–92 years) and 63.6 years (range: 24–96 years) (p < 0.001); 164.9 cm (range: 130–192 cm) and 171.9 cm (range: 149–200 cm) (p < 0.001); 73.6 kg (range: 38–170 kg) and 85 kg (range: 42–138 kg) (p < 0.001). The mean age, mean height, and mean weight of the Hydra/Hydra-fix cohorts were, respectively, 69.3 years (range: 34–90 years) and 67.7 years (range: 40–90 years) (p < 0.001); 164.6 cm (range: 140–195 cm) and 168.7 cm (range: 140–198 cm) (p < 0.001); 73 kg (range: 33–150 kg) and 81.4 kg (range: 45–125 kg) (p < 0.001). The mean age, mean height, and mean weight of the Recta/Recta-fix cohorts were, respectively, 69.1 years (range: 35–90 years) and 67.7 years (range: 44–85 years) (p = 0.087); 166.6 cm (range: 140–190 cm) and 167.9 cm (range: 140–190 cm) (p = 0.009); 75.6 kg (range: 42–120 kg) and 79.9 kg (range: 47–149 kg) (p < 0.001). The distribution per age decade and body mass index (BMI) class are detailed in the table: there was a significant difference in terms of age decade for Apta/Apta-fix and Hydra/Hydra-fix cohorts (p < 0.001) and for all the three pairs in terms of BMI class (Apta/Apta-fix, p < 0.001; Hydra/Hydra-fix, p < 0.001; Recta/Recta-fix, p = 0.04) (Table 1). The two pairs of cohorts Apta/Apta-fix and Hydra/Hydra-fix were not comparable in terms of head size (higher 36 mm rates in the ST cohorts, p < 0.001) and cup size (bigger sockets in ST cohorts, p < 0.001). Recta/Recta-fix cohorts were comparable in terms of head size (p = 0.091) and cup size (p = 0.316). The three pairs of stems were not comparable in terms of stem size (larger sizes in DT cohorts, p < 0.001).

For each stem design, the two curves were fully reliable at 5 years (> 10% of the implants at risk). At a mid-term follow-up, ST and DT versions achieved comparable survival rates (Apta/Apta-fix: p = 0.076; Hydra/Hydra-fix: p = 0.319; Recta/Recta-fix: p = 0.616) (Figs. 3, 4, 5). For each stem design, the HR adjusted for age (categorical variable, > 65 years or ≤ 65 years) and sex showed that the two versions were not different in terms of revisions (Apta/Apta-fix: p = 0.084; Hydra/Hydra-fix: p = 0.308; Recta/Recta-fix: p = 0.729). For each stem design, the reasons for revision of the two versions are reported in Table 2.

Kaplan-Meier curves of the Apta (in red) and Apta-fix (in blue) stems: the two versions achieved similar survival rates when the endpoint was revision for any reason. y-axis, percentage of survival implants; x-axis, years

Kaplan–Meier curves of the Hydra (in red) and Hydra-fix (in blue) stems: the two versions achieved similar survival rates when the endpoint was revision for any reason. y-axis, percentage of survival implants; x-axis, years

Kaplan–Meier curves of the Recta (in red) and Recta-fix (in blue) stems: the two versions achieved similar survival rates when the endpoint was revision for any reason. y-axis, percentage of survival implants; x-axis, years

When stem-focused endpoints were adopted, ST and DT versions of the three stems achieved similar survival rates at 5 years (Apta/Apta-fix: p = 0.710; Hydra/Hydra-fix: p = 0.784; Recta/Recta-fix: p = 0.983). For each stem, the HR adjusted for age and sex showed that the two versions were comparable in terms of revisions for stem and neck failures (Apta/Apta-fix: p = 0.647; Hydra/Hydra-fix: p = 0.858; Recta/Recta-fix: p = 0.787). Three neck breakages occurred (0.0007% of all the modular implants), two with Apta and one with Recta stems (all with 36 mm heads). The two neck failures in the Apta cohort occurred in obese men (BMI 31 and 36 kg/m2; weight > 90 kg); the patients were 67 years old and 72 years old. The breakages occurred 4.94 and 3.95 years after the first implant. Both the failures underwent neck exchange with no stem removal. The neck failure in the Recta cohort occurred in a 52-year-old man (BMI 27 kg/m2, weight 90 kg), 7.67 years after the first implant. In this case, a stem revision was performed.