Do Clinical Parameters Reflect Local Bone Metabolism in Heterotopic Ossification After Septic or Aseptic THA?

Heterotopic ossification (HO) is a common complication after THA. Although current research primarily focuses on treatment and prevention, little is known about the local bone metabolism of HO and clinical contributing factors.

We aimed to assess bone remodeling processes in HO using histomorphometry, focusing on the effects of inflammation and prior NSAID treatment. Specifically, we asked: (1) Are HO specimens taken from patients with periprosthetic joint infection (PJI) more likely to exhibit active bone modeling and remodeling than specimens taken at the time of revision from patients without infection? (2) Do clinical or inflammatory serum and synovial parameters reflect the microstructure of and remodeling in both HO entities? (3) Is NSAID treatment before revision surgery associated with altered local bone mineralization or remodeling properties?

Between June 2021 and May 2022, we screened 395 patients undergoing revision THA at two tertiary centers in Germany. Of those, we considered all patients with radiographic HO as potentially eligible. Based on that, 21% (83 of 395) were eligible; a further 43 were excluded because of an inability to remove the implant intraoperatively (16 patients), insufficient material (11), comorbidities with a major effect on bone metabolism (10), or bone-specific drugs (six), leaving 10% (40) for analysis in this retrospective, comparative study. HO specimens were collected during aseptic (25 patients: 18 male, seven female, mean age 70 ± 11 years, mean BMI 29 ± 4 kg/m 2 ) and septic (15 patients: 11 male, four female, mean age 69 ± 9 years, mean BMI 32 ± 9 kg/m 2 ) revision THA at a mean of 6 ± 7 years after primary implantation and a mean age of 70 ± 9 years at revision. Septic origin (PJI) was diagnosed based on the 2018 International Consensus Meeting criteria, through a preoperative assessment of serum and synovial parameters. To specify the local bone microstructure, ossification, and cellular bone turnover, we analyzed HO specimens using micro-CT and histomorphometry on undecalcified sections. Data were compared with those of controls, taken from femoral neck trabecular bone (10 patients: five female, five male, mean age 75 ± 6 years, mean BMI 28 ± 4 kg/m 2 ) and osteophytes (10 patients: five female, five male, mean age 70 ± 10 years, mean BMI 29 ± 7 kg/m 2 ). The time between primary implantation and revision (time in situ), HO severity based on the Brooker classification, and serum and synovial markers were correlated with HO microstructure and parameters of cellular bone turnover. In a subgroup of specimens of patients with NSAID treatment before revision, osteoid and bone turnover indices were evaluated and compared a matched cohort of specimens from patients without prior NSAID treatment.

Patients with aseptic and septic HO presented with a higher bone volume (BV/TV; aseptic: 0.41 ± 0.15, mean difference 0.20 [95% CI 0.07 to 0.32]; septic: 0.43 ± 0.15, mean difference 0.22 [95% CI 0.08 to 0.36]; femoral neck: 0.21 ± 0.04; both p < 0.001), lower bone mineral density (aseptic: 809 ± 66 mg HA/cm 3 , mean difference -91 mg HA/cm 3 [95% CI -144 to -38]; septic: 789 ± 44 mg HA/cm 3 , mean difference -111 mg HA/cm 3 [95% CI -169 to -53]; femoral neck: 899 ± 20 mg HA/cm 3 ; both p < 0.001), and ongoing bone modeling with endochondral ossification and a higher proportion of woven, immature bone (aseptic: 25% ± 17%, mean difference 25% [95% CI 9% to 41%]; septic: 37% ± 23%, mean difference 36% [95% CI 19% to 54%]; femoral neck: 0.4% ± 0.5%; both p < 0.001) compared with femoral neck specimens. Moreover, bone surfaces were characterized by increased osteoblast and osteoclast indices in both aseptic and septic HO, although a higher density of osteocytes was detected exclusively in septic HO (aseptic: 158 ± 56 1/mm 2 versus septic: 272 ± 48 1/mm 2 , mean difference 114 1/mm 2 [95% CI 65 to 162]; p < 0.001). Compared with osteophytes, microstructure and turnover indices were largely similar in HO. The Brooker class was not associated with any local bone metabolism parameters. The time in situ was negatively associated with bone turnover in aseptic HO specimens (osteoblast surface per bone surface: r = -0.46; p = 0.01; osteoclast surface per bone surface: r = -0.56; p = 0.003). Serum or synovial inflammatory markers were not correlated with local bone turnover in septic HO. Specimens of patients with NSAID treatment before revision surgery had a higher osteoid thickness (10.1 ± 2.1 µm versus 5.5 ± 2.6 µm, mean difference -4.7 µm [95% CI -7.4 to -2.0]; p = 0.001), but there was no difference in other osteoid, structural, or cellular parameters.

Aseptic and septic HO share phenotypic characteristics in terms of the sustained increase in bone metabolism, although differences in osteocyte and adipocyte numbers suggest distinct homeostatic mechanisms. These results suggest persistent bone modeling or remodeling, with osteoblast and osteoclast indices showing a moderate decline with the time in situ in aseptic HO. Future studies should use longitudinal study designs to correlate our findings with clinical outcomes (such as HO growth or recurrence). In addition, the molecular mechanisms of bone cell involvement during HO formation and growth should be further investigated, which may allow specific therapeutic and preventive interventions.

To our knowledge, our study is the first to systematically investigate histomorphometric bone metabolism parameters in patients with HO after THA, providing a clinical reference for evaluating modeling and remodeling activity. Routine clinical, serum, and synovial markers are not useful for inferring local bone metabolism.