Application of POSSUM and P-POSSUM scores in the risk assessment of elderly hip fracture surgery: systematic review and meta-analysis

Hip fractures mainly refer to femoral neck fractures and intertrochanteric fractures. Gullberg et al. [1] predict that the global number of hip fractures will reach 2.6 million and 4.5 million by 2025 and 2050. With the aging of the global population, the incidence of elderly hip fractures in developed countries can reach 350/100,000[2], becoming one of the most common diseases in joint orthopedics. Generally, the main treatment for hip fractures is surgery, however elderly patients with hip fractures often have many basic diseases when they are admitted to the hospital. George et al. [3] found that when elderly patients with hip fractures had a variety of other diseases, the risk of postoperative morbidity and mortality increased. Their data showed that 20% of the patients had postoperative complications, and the 30-day death rate of such patients was 9%. In addition, previous studies have also shown that the ageing of patients, cognitive dysfunction and other body aging processes will also increase the risk of postoperative complications or death in patients with hip fractures [4‐6]. In actual clinical work, active preoperative preparation will significantly reduce the adverse events of patients during the perioperative period. Surgical risk prediction models are a very valuable tool for surgeons because these tools allow surgeons to assess and prevent patients’ perioperative events in advance, so as to make the best decisions to optimize resources and improve the quality of care for patients [7, 8].

Many prediction models have emerged in recent years, but their specific application value is almost limited. The Physiological and Operative Severity Score for enUmeration of mortality and morbidity (POSSUM) was first proposed by Copeland et al. [9] in 1991 to evaluate the prognosis of patients. Subsequently, Whiteley et al. [10] found that the POSSUM scoring system would overestimate the postoperative mortality rate, and after simplifying the exponential analysis technology in the POSSUM scoring system, a P-POSSUM scoring formula that was more suitable for surgical patients was obtained. Currently, the APACHEI, APACHE II, POSSUM, ASA and NHFS scoring systems are generally used internationally to assess the perioperative risk of surgical patients, and the research by de Cássia Braga Ribeiro et al. [11] believes that the POSSUM scoring system has the most application value in the risk assessment of the perioperative period. Some scholars believe that the Surgical Risk Scale (SRS) is more accurate than the POSSUM scoring system for surgical risk assessment, has the advantages of simple procedures, and can obtain the predictive indicators before surgery [12], but in disciplines of general surgery, vascular surgery and esophageal surgery, the POSSUM scoring system is still widely used and recognized by surgeons [13‐17].

Since Mohamed et al. [18] analyzed 2326 orthopedic cases in 2002 and believed that the POSSUM formula can be directly used to predict postoperative complications and death risks in orthopedic patients, more and more reports have appeared on the application of POSSUM and P-POSSUM scores in hip fracture surgery. Yet, there were inconsistencies in the related reports, some reports believed that the POSSUM scoring system could not accurately predict postoperative morbidity and mortality, but other reports supported this model [19‐22]. Therefore, the purpose of this study is to conduct a meta-analysis on the application value of POSSUM and P-POSSUM in the risk assessment of hip fracture surgery, in order to guide the orthopedic surgeon in the evaluation of postoperative risk events and the choice of surgical benefits.

According to research reports that can be retrieved so far, this study is the first comprehensive analysis of the application of POSSUM and P-POSSUM scoring systems to elderly hip fractures. In terms of the results of this study, the POSSUM scoring system can accurately predicted the postoperative morbidity (RR = 1.07, 95% CI 0.93–1.24), and compared with POSSUM over-predicted the mortality (RR = 1.93, 95% CI 1.21–3.08), P-POSSUM can accurately predict the postoperative mortality of elderly patients with hip fractures (RR = 1.15, 95% CI 0.89–1.50).

In a meta-analysis report on the application value of POSSUM scoring system in hepatobiliary and pancreatic surgery by Chen et al. [33], they believe that because the early POSSUM scoring model is not suitable for current surgical operations, it led to POSSUM’s over-prediction of postoperative morbidity and mortality. Similarly, we also believe that the same situation may exist here. The POSSUM scoring system proposed by Copeland et al. in 1991 was originally used for general surgery patients. It was used in orthopedics after Mohamed et al. [18] modified its surgical scoring table according to the characteristics of orthopedic surgery in 2002. In the past 18 years, surgical methods and instruments have been greatly improved, and the nursing teams in related departments have become more professional. Previous improvements according to the characteristics of orthopedic surgery are no longer applicable to the current postoperative risk assessment. And, with the development of surgical technology, the impact of surgical intervention on the prognosis of patients is getting smaller and smaller. Combining the characteristics of high incidence of hip fracture in the elderly patients and having many underlying diseases in this group of people, it is not enough to modify the surgical severity score scale alone, and the physiological score scale should also be specifically modified. Orthopedic patients, especially those with hip fractures, have a tendency to stay in bed for a long time, which leads to a high risk of venous thrombosis. The physiology score scales, such as coagulation function, D-dimer, and other auxiliary examinations such as neck and lower extremity vascular color Doppler ultrasound to assess the patient's vascular condition, have not been reflected in the score.

Secondly, some studies reported that the difference in the prediction by POSSUM on the morbidity and mortality may be caused by the difference in the medical technology level, population characteristics, and the composition ratio of patient operation types in the non-originating country of the system from the originating country of the system [12]. In the process of exploring the source of heterogeneity of combined results in this study, we found that the type of hip fracture, the choice of surgical methods, and the distribution of patient age may all have an impact on the prediction accuracy of the POSSUM scoring system. Because the treatment methods of different fracture types are often different, the prognosis is naturally very different; and advanced surgical methods and young patients show a lower risk of postoperative death and complications. In addition, the POSSUM score of elderly patients may have false score values. For example, a study by Ramanathan et al. [21] found that elderly people older than 80 years old are abnormal compared to normal people because the relevant test results are "normal" for themselves, which leads to the inability of the POSSUM score to accurately predict the postoperative complications rate and mortality of this group of people. That is, POSSUM gives an excessively high score for the "normal" physiological examination results that can maintain the physiological function of this group of people. Therefore, this group population may not be suitable for POSSUM scoring system, a two-component scoring model of preoperative physiology score and surgical severity score. We believe that the above problems may be solved by risk grouping based on the size of the POSUUM score value. From previous reports, it can be seen that the error of POSSUM's over-predicting postoperative death events mainly comes from the low-risk group. For example, in the reports of Whiteley et al. [10] and Prythech et al. [34], The predicted value of death event in the low-risk group was 6 times and 7 times the actual observation value, and POSSUM performed well in the high-risk group. Therefore, we believe that it is necessary to improve the predictive ability of the POSSUM scoring system in the complication rate and mortality of patients with hip fractures, and it is very necessary to predict the risk of patients when using the scoring system.

It is worth raising that, according to previous reports in the literature, the overall prediction of the POSSUM scoring system for postoperative patients was consistent with actual observations, but its prediction for younger patients and patients undergoing elective surgery was found to be higher than the actual outcome in the subgroup [35], so this seems to suggest that the elderly population and critically ill patients requiring emergency surgery may be the priority groups. Evidence for this view is also provided by a recently retrospective cohort study, which showed that the P-POSSUM score showed good predictive power for postoperative mortality in COVID-19 positive patients undergoing emergency general surgery [36]. The scoring system is not applicable to children because the physiological scoring indicators in the scoring system are those of adults. In addition, the definition of postoperative complications is not completely clear, especially for those with preoperative underlying disease, so a clear definition of postoperative complications is necessary. Finally, according to the regulations of the POSSUM scoring system when it was established, the physiological indicators of the system should preferably be data within 1 day before surgery, and POSSUM is best for predicting complications and mortality within 30 days after surgery. And regarding the relevant data in the surgical scoring, such as blood loss and operative time, they need to be recorded accurately by the assessor.

One of the shortcomings of this study is that although our study was analysed in subgroups, we did not find a major source of heterogeneity in the POSSUM-predicted postoperative mortality pairs. We suspect that this may be due to differences in the cause of the fracture, the physical impact of the fracture itself, the treatment, the distribution of the population, and the level of medical care, but we were unable to draw firm conclusions. In addition, this study also failed to collect relevant unpublished data.

In all, the current data analysis shows that the POSSUM scoring system can predict the postoperative morbidity of elderly hip fracture patients. POSSUM's prediction of mortality is affected by the type of fracture and size of sample. Compared with POSSUM, P-POSSUM can accurately predict the postoperative mortality of patients with hip fractures. The application of the POSSUM scoring system in orthopedics needs to be further improved according to the characteristics of orthopedic patients and surgery, as well as needing more multi-center and large-sample prospective studies so that it can be more applicable to orthopedics.