In our study, the most important finding was that the specificity of mid-piece FPR for PJI was 96.5%, and the PPV was 94.1%. When combined with elevated serum CRP / ESR, we doubted whether there was an infection. The results of this study suggest that FPR has certain diagnostic value in hip joint prosthesis infection.
To date, multiple research institutions and orthopaedic associations throughout the world have released a number of PJI diagnostic consensuses and still make every effort to improve criteria, which are focused on culture and involve laboratory tests of joint fluid/periprosthetic tissues/retrieved implants(6, 14, 15). For conventional radiography, limited diagnostic research has been published for PJI, and it is more commonly used in the treatment stage to provide some degree of confidence in planning the extent of bone resection needed during resection arthroplasty(16, 17). The presence of osteolytic lesions and loosening and effusion of periprosthetic soft tissues indicate the presence of infection around the prosthesis(14, 18, 19). Esposito et al. found that bone erosion and new bone formation may indicate the presence of PJI in 3 to 6 months after surgery(20). Tigges et al. suggest that the diagnostic value of imaging can be considered very limited because no obvious abnormality was observed in most of the radiography imaging of 20 HPJI patients retrospectively(21). Kapadia et al. believe that the presence of wide radiolucency at the interface between bone and implants (cement-bone interface) and the presence of bone destruction may also indicate infection but have little practical effect on the diagnosis of disease(22). Periosteal response exists as a manifestation of periosteal inflammatory changes. Cytesval et al. have shown that imaging changes of bone abnormalities are significant in the diagnosis of hip joint prosthesis implantation, but the diagnostic value of joint infections remains uncertain(23). However, encouragingly, our study found that the periosteal reaction has a high specificity in HPJI, especially when this result is combined with a positive result of laboratory serology, which can obtain a very reliable diagnostic specificity and PPV. It is of great help in the early diagnosis of joint infection in clinics.
The periosteal reaction is a process in which a patient's periosteum is subjected to various stimuli, presenting as activity of osteoblasts in the inner layer of the bone increases and growth of periosteal new bone. There are numerous kinds of morphologies under the influence of different external stimuli, considering as an imaging feature of bone or surrounding and playing an extremely important role in identifying the location, characterization, and extent of lesions. All patients in our study were classified as chronic infections according to the Tsukayama classification. A number of studies in recent years have shown that chronic infections are mostly related to biofilm formation and immune escape(24, 25), and long-lasting local inflammation causes periosteal hyperplasia of the inner and outer cortical bone presented as layered or dense periosteal appearance(26, 27). In addition, studies in the infection group showed that not all patients had a positive periosteal reaction, which was similar to previous studies(21, 23, 28). X-ray imaging results can only reflect mineralization or periosteal osteogenesis with calcium salt deposition. Some studies have shown that periosteal mineralization takes approximately 10–21 days(29), which is why the neonatal periosteal reaction would be easily ignored by doctors. Therefore, patients with acute prosthetic infection in the acute stage are mostly normal on imaging or only have displacement of the prosthesis. In general, the periosteal reaction caused by infection is mostly attributed to pus infiltration under the periosteal stimulation(10). The severity of the infection is related to the bacterial virulence and local immune microenvironment.
On the other hand, the overall positive rate of periosteal reaction in patients was low, only approximately 33%, but some patients in the non-infected group also observed its existence, which led to a decrease in the sensitivity and specificity. However, through demographic characteristics and imaging tests of the cases, we found that the FPR-positive cases in the non-infected group were all overweight women approximately 50 years old, except for one case (Fig. 1c), and most of the reactions were located at the end-piece of the prosthesis characterized by expansive hyperplasia. Pawel Szulc et al. showed that cortical bone loss was noticeable in perimenopausal women, while it was relatively slow in men(30). The bone-prosthesis-cement interface is formed at the end of the prosthesis after joint replacement, which belongs to the stress concentration point. By strong mechanical stimulation, local microfracture arises, and the periosteal reaction is obviously concentrated without spreading upward to the upper and middle parts of the prosthesis(31, 32). In addition, the routine laboratory tests in the non-infected group did not show significant positive changes, which can be ruled out by combined diagnosis.
There are several limitations in our research. First, our research is limited by the inherent shortcomings of retrospective studies. It is a single-centre case review, so the number of patients included in the study is relatively small. Although we set up a control group for comparative studies, the overall positive rate of periosteum response is not high. The study set up a double-blind imaging test to collect image data, but it could still not avoid the existence of selective migration. Second, although X-ray imaging has better spatial resolution, it is more inexpensive and much more convenient. Compared with CT scan, imaging still has many disadvantages, such as unclear resolution and easy missed diagnosis of periosteal reaction. Studies have shown that multi-detector CT can better compensate for the above disadvantages(33). In the future, research directions may need to use more advantageous imaging methods to assist in diagnosis.