ESTIMATION OF CLINICAL USE OF PREDICTIVE MODEL OF RISK OF COMPLICATIONS FOR EFFICIENT SURGICAL TREATMENT OF PATIENTS WITH PROXIMAL FEMORAL FRACTURES Ustyantsev D.D., Milyukov A.Yu., Agadzhanyan V.V., Gilev Ya.Kh., Vlasov S.V.
Regional Clinical Center of Miners’ Health Protection, Leninsk-Kuznetsky, Russia,
Tsyvyan Novosibirsk Research Institute of Traumatology and Orthopedics, Novosibirsk, Russia
The population of older and senile persons has been increasing in the world [1, 2]. As result, the amount of hospital admissions of older patients with traumatic injuries has been increasing [3]. Proximal femoral fractures are annually registered in 2 million people each year [4, 5]. The International Fund of Osteoporosis predicts about 6 million patients with proximal femoral fractures in 2050 [3, 4, 5].
In Russia, this diagnosis is confirmed in 100-150 thousand persons per 100,000 of population, with increasing rate [5, 6]. Mainly, the amount of patients with proximal femoral fractures has been growing simultaneously with increasing amount of older patients since more than a half of such injuries happens in persons older 60 [6, 7, 8].
Patients older 65 are characterized with high risk of complications in postsurgical period [8, 9], with higher post-injury mortality as compared to younger people [9, 10]. It is associated with comorbid diseases [11, 12] and decreasing physiologic reserves of the body [13].
Considering the increasing risk of complications and mortality in older patients with proximal femoral fractures, it is obligatory to discuss the issues of surgical treatment with consideration of comorbid status [6, 8, 11]. Although some well-known indicators of unfavorable prognosis are available for this age group [9, 10, 12, 13], there are not any generally accepted predictive criteria in relation to clinical estimation of the risk of postsurgical complications at the background of concurrent diseases in patients with proximal femoral fractures. Therefore, it is necessary to develop a prediction model, which can be used during treatment of such patients.
The previous studies present our prediction model for estimation of the risk of postsurgical complications in proximal femoral fractures with consideration of 5 parameters (gender, age, a comorbidity category, ASA condition severity class, a fracture type according to the nomogram) [14, 15]. Subsequently, we offered a surgical technique for patients with proximal femoral fractures as a tool for decreasing the potential and actual risk of postsurgical complications [16]. As result, the next stage should include the objective comparison and analysis of clinical results after surgical treatment of patients with proximal femoral fractures with consideration of possible complications.
Objective – to estimate the clinical use of predictive model of risk of postsurgical complications for efficient surgical treatment (osteosynthesis and primary total endoprosthetics) for patients with proximal femoral fractures by means of analysis of complications, treatment duration, functional results and mortality.
PATIENTS AND METHODS
The study was conducted in compliance with World Medical Association Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects, 2013, and the Rules for Clinical Practice in the Russian Federation (the Order by RF Health Ministry, 19 June 2003, No.266), with written consent for participation in the study and the approval from the local ethical committee of the center. 90 patients (the main group) with proximal femoral fractures were operated in Regional Clinical Center of Miners’ Health Protection from January 2017 till December 2018. The control group included 145 patients with proximal femoral fractures who were in the center in the period of retrospective studies from January 2013 till December 2016.
The inclusion criteria were the age ≥ 18, ISS ≤ 15 [17], no transfer to other clinics, more than 24 hours of stay in the clinic).
The table 1 presents the findings of demographic, clinical and physiological parameters in patients with proximal femoral fractures after hospital admission.
The following parameters were considered for each patient: age ((18-64 years) – young; (65+) – older), gender, comorbid status (0, 1-2, 3+ concurrent diseases), ASA severity class [18], a fracture type, a surgery type (internal fixation or joint replacement), hospital stay before (1 day or 2, 3 or more days) and after surgery.
The objective estimation of adequate preparation for surgery was realized with our prediction model of postsurgical complications with use of nomograms for men and women [14, 15]. The predictive risk of complications was estimated in relation to increasing age, comorbidity categories, ASA severity class, a fracture type (intraarticular medial fractures or femoral neck fractures: subcapital, transcervical, basicervical; extraarticular lateral or trochanteric fractures – transtrochanteric, subtrochanteric) for men and women [14].
In compliance with ICD-10, proximal femoral fractures were verified on the basis of complaints, physical examination and X-ray examination in two standard planes. A type of proximal femoral fracture was confirmed according to the classification by A.V. Kaplan [19].
The patients with proximal femoral fractures were distributed into groups of low (< 10 %), average (10-30 %) and high risk (> 30 %) of postsurgical complications on the basis of our model [14].
The surgical treatment included fixation with PFN, cannulated screws, and primary total hip replacement according to the modern guidelines [20]. A fixation type (screws or PFN) was selected according to fracture pattern.
A possibility of surgery (fixation or endoprosthetics) for proximal femoral fractures was estimated with consideration of risks of postsurgical complications relating to the age, concurrent pathology, ASA severity class according to our technique [16]. If risks were estimated as moderate, the surgical treatment was conducted.
Mobilization of patients was initiated in the first day after surgery. The walking frame and crutches were used. The dosed (30 % of body weight) load to the operated extremity was allowed in dependence on individual level of rehabilitation and intensity of pain.
In the postsurgical period, the types and amount of complications, treatment duration and mortality were registered.
Estimation of functional results of surgical treatment of patients with proximal femoral fractures was conducted 1 year after surgery with use of the descriptive scale from American Academy of Orthopedic Surgeons Assessment, developed by R.A. Goodwin in 1968 [21], and Harris Evaluation System of the Hip [22]. These techniques allow estimating the clinical outcomes after fixation of proximal femoral fractures and hip replacement.
Goodwin score for estimation of hip surgery outcomes includes the qualitative estimation of criteria (pain, volume of motions, walking), with estimated results as fine, good, satisfactory or unsatisfactory [21].
Harris Evaluation System of the Hip supposes the estimation of four categories: pain, function, deformation, range of motions. Each category is characterized by specific number of points. The maximal number of points is 100. 90-100 points present excellent function, 80-89 – good, 70-79 – satisfactory, < 70 – unsatisfactory [22].
Statistical analysis
The statistical analysis of the data was conducted with IBM SPSS Statistics 21 (Statistical Product and Service Solutions – SPSS).
The qualitative signs are presented as absolute and relative (%) values. The quantitative variables are presented as mean arithmetic (M) and standard deviation (SD) of mean arithmetic in view of Me (LQ-UQ), where Me – median, (LQ-UQ) – interquartile range (IQR) (LQ – 25 %, UQ – 75 % quartiles). Depending on type of distribution of variables, Student’s test or Mann-Whitney’s U-test were used for estimation of reliability of differences in the groups. χ2 (chi-square) test was used for estimation of significance of differences in incidence of the studied values. The critical level of significance (α) during testing the statistical hypotheses was 0.05. Differences were considered as statistically significant with p < 0.05.
RESULTS
The table 1 shows the high degree of matching between the characteristics of demographic, clinical and physiological parameters in patients with proximal femoral fractures at admission in the main (n = 90) and comparison (n = 145) groups. Proximal femoral fractures were presented by intraarticular medial or femoral neck fractures in 76 %, and by extraarticular lateral or trochanteric fractures in 24 % in the main and comparison groups (the table 1).
Table 1
Characteristics of demographic, clinical and physiological parameters in patients with proximal femoral fractures at admission to the clinic
|
Parameters
|
Main group (n = 90), abs. (%) 1
|
Comparison group (n = 145), abs. (%) 1
|
p |
|
Age: young age (18-64 years)
older age (65 +)
|
19 (21) 71 (79)
|
30 (20.7) 115 (79.3)
|
0.833 0.878 |
|
Gender: |
|
|
|
|
Men Women |
51 (57) 39 (43) |
84 (57.9) 61 (42.1) |
0.652 0.69 |
|
Injury severity (ISS)2, M (SD), points: |
|
|
|
|
|
12 (9.1) |
13 (8.9) |
0.873 |
|
Fracture type3: |
|
|
|
|
Medial fractures of femoral neck (intraarticular): |
69 (76) |
111 (76) |
0.971 |
|
subcapital |
21 (23) |
32 (22) |
1.0 |
|
transcervical |
29 (32) |
47 (32) |
1.0 |
|
basicervical |
19 (21) |
32 (22) |
0.983 |
|
|
|
|
|
|
Lateral trochanteric fractures (extraarticular): |
21 (24) |
34 (24) |
0.863 |
|
transtrochanteric |
14 (16) |
24 (17) |
0.852 |
|
subtrochanteric |
7 (8) |
10 (7) |
0.806 |
|
Comorbidity (concurrent diseases before injury): |
|
|
|
|
no concurrent diseases (0) |
7 (7.7)
|
10 (6.9)
|
1.0 |
|
1-2 concurrent diseases |
50 (55.6) |
80 (55.2) |
0.724 |
|
3+ concurrent diseases |
33 (36.7) |
55 (37.9) |
0.733 |
|
ASA severity class4:
1 |
7 (7.8) |
13 (9) |
0.875 |
|
2 |
18 (20) |
24 (16.6) |
1.0 |
|
3 |
41 (45.6) |
70 (48.2) |
0.830 |
|
4 |
24 (26.6) |
38 (26.2) |
1.0 |
|
|
|
|
|
Note:
1 percentage correlation was calculated with consideration of all patients in the groups.
2 ISS – Injury Severity Score [17].
3 Fracture type according to classification
by A.V. Kaplan [19].
4 Severity class according to classification of objective status of American Society of Anesthesiologists (ASA) [18]. M (SD) – mean (standard deviation); p – probability of absence of intergroup differences.
Initially, our prediction model was used for estimation of the risk of possible postsurgical complications in the main group [14]. After distribution of the patients (treated from January 2017 till December 2018) into the risk groups, 4 patients (4.4 %) were in the low risk group (< 10 %); the average risk group (10-30 %) included 56 patients (62.2 %); 30 patients (33.4 %) were in the high risk group (> 30 %) (the table 2).
Table 2
Comparative characteristics of demographic predictive risk of complications in patients with proximal femoral fractures in the studied groups.
|
Parameters
|
Main group (n = 90), abs. (%) 1
|
Comparison group (n = 145), abs. (%) 1
|
p |
|
Predictive risk of complications2 at admission to hospital, n (%): |
|
|
|
|
Low (0-10 %) |
4 (4.4) |
10 (6.9) |
0.650 |
|
Middle (11-30 %) |
56 (62.2) |
80 (55.2) |
0.740 |
|
High (> 30 %) |
30 (33.4) |
55 (37.9) |
0.850 |
|
Predictive risk of complications2 before surgery (recurrent estimation), n (%):
|
|
|
|
|
Low (0-10 %) |
10 (11) |
|
|
|
Middle (11-30 %) |
58 (64) |
|
|
|
High (> 30 %) |
22 (25) |
|
|
|
|
|
|
|
|
Surgery type, abs. (%): |
|
|
|
|
Osteosynthesis: |
35 (28.9) |
55 (28.9) |
1.0 |
|
cannulated screws |
20 |
25 |
|
|
PFN |
15 |
30 |
|
|
Total hip replacement |
55 (71.1) |
90 (71.1) |
1.0 |
|
|
|
|
|
|
Days before surgery: |
|
|
|
|
1 |
4 (4.4) |
10 (6.9) |
0.74 |
|
2 |
10 (11.1) |
27 (18.6) |
0.28 |
|
3+ |
76 (84.5) |
108 (74.5) |
0.04 |
|
Type of complication before surgery1, n (%): |
8 (8.8) |
22 (15.2) |
0.009
|
|
Complications after fixation with screws and PFN: |
4 (4.4) |
12 (8.3) |
0.007 |
|
|
|
|
|
|
|
|
|
|
|
Migration of metal constructs |
1 (1.1) |
1 (0.6) |
0.93 |
|
Thrombosis of lower extremity veins |
3 (3.3) |
10 (6.8) |
0.008 |
|
Contact dermatitis
|
|
1 (0.6) |
|
|
Complications after total hip replacement: |
4 (4.4) |
10 (6.9) |
0.008 |
|
Local infection in surgical site |
|
1 (0.67) |
|
|
Endoprosthesis head displacement |
|
1 (0.67) |
|
|
Thrombophlebitis and thrombosis of lower extremity veins |
4 (4.4) |
7 (4.89) |
0.04 |
|
Contact dermatitis |
|
1 (0.67) |
|
|
|
|
|
|
|
ICU stay, Me (IQR), days: |
0.7 (0.6 – 1.0) |
0.9 (0.6 – 1.2) |
0.11 |
|
Hospital stay after surgery, Me (IQR), days: |
11.3 (9.2 – 15.0) |
14.5 (9.2 – 23.0) |
0.04 |
Note:
1 percentage correlation was calculated with consideration of all patients in the groups.
2 Predictive risk of postsurgical complications (low (0-10 %), middle (11-30 %), high (> 30 %)) was calculated with use of the model considering the age, gender, comorbidity status, ASA severity class and fracture type with nomogram for men and women individually.
The comparison of values of initial estimation of predictive risk of complications at admission did not find any significant differences in distribution of patients into the risk group in the main and comparison groups (the table 2).
The patients of the main group with high risks of surgical intervention (n = 30) relating to concurrent pathology and severity of condition received the conservative symptomatic treatment of cardiovascular, therapeutic and surgical diseases under supervision of cardiologist, therapeutist and surgeon within 7 days (on average, 6.8 (2.40 days before surgery)). After correction of somatic pathology, the recurrent estimation of the risk of postsurgical complications was conducted in the main group.
The recurrent presurgical use of the prediction model of the risk of complications with use of the nomogram showed the low, average and high risk of postsurgical complications correspondingly in 10 (11 %), 58 (64 %) and 22 (25 %) patients with proximal femoral fractures (the table 2). In the main group, the number of patients with high risk of postsurgical complications decreased 1.34 times (χ2 = 36.2, p < 0.01) in relation to the initial value by means of redistribution into the groups of average and low risk (2.5-fold increase in the number of patients with low risk as compared to the initial estimation, χ2 = 23.2, ð < 0.01) (table 2).
Our method for selection of management techniques for proximal femoral fractures [16] was taken as the basis for the surgical management algorithm with use of the prediction model of the risk of postsurgical complications (Fig. 1).
Figure 1
Surgical treatment algorithm (osteosynthesis or primary total hip replacement) for patients with proximal femoral fractures with consideration of predictive risk of postsurgical complications.
The potential risk of complications (more or less 30 %) was estimated with the nomogram (combination of parameters) with consideration of the age (young (18-64 years) and older (65+)), a comorbidity category (0 – absence, 1-2, 3 and more concurrent diseases), a severity class according to ASA classification of objective status of patient, a type of fracture in men and women separately [14, 15].
If ASA severity class 4 with potential risk of complications > 30 % was found, surgical interventions were not performed on the admission day, and the patients were transferred to the profile units (cardiology, neurology, therapy and surgery) for condition stabilization. Subsequently, the potential risk of complications was estimated again, and a possibility of surgical treatment was estimated. For ASA severity class 3 with potential risk of complications < 30 %, low invasive fixation was conducted for proximal femoral fractures in medial or lateral region. A fixation method was selected according to a fracture type (screws or PFN): closed reposition and cannulated screws for medial fractures; closed reposition and intramedullary fixation with PFN for lateral injuries.
For ASA severity classes 1 or 2 with potential risk of complications < 30 %, the principle of unlimited (required) degree of surgical invasion was used. For medial fractures (subcapital and transcervical), total hip replacement was carried out.
For basicervical fractures of femoral neck, the quantitative analysis of X-ray images was conducted for estimation of peripheral index of structural changes (cortical index) with measurement of thickness of cortical layer of the femoral bone according to Barnett-Nordin [23]. Total hip replacement was conducted if cortical index was < 54 %. If cortical index was ≥ 54 %, the patient's physical activity was measured with Rivermead Mobility Index (RMI) [24]. For high mobility index (7-15 points), closed reposition and intramedullary locked fixation with PFN were performed. For low level of physical activity (≤ 7 points), closed reposition and fixation with cannulated screws were performed.
For lateral proximal femoral injuries, namely for subtrochanteric fractures, closed reposition and locked fixation with PFN were used. For trochanteric fractures, the intensity of peripheral structural changes was measured with cortical index according to Barnett-Nordin. If cortical index was ≥ 54 %, closed reposition and PFN fixation were conducted. If Barnett-Nordin index was < 54 %, the condition of bone tissue was estimated with Singh index [25] for qualitative estimation of intensity of osteoporosis in the femoral head and neck, and in greater trochanter with 7-point scale. If Singh index was ≤ 2 (i.e. full disappearance of archy trabecules), total hip replacement was conducted.
The table 2 presents the types of surgeries and their amount according to types of proximal femoral fractures in the patients of the study groups.
Primary total hip replacement and osteosynthesis were conducted for 71.1 % and 28.9 % correspondingly. It is necessary to note that only 25 % of the patients of the main group showed high risk (> 30 %) of postsurgical complications in comparison with 37.9 % of patients in the comparison group (p = 0.03). It is associated with adequate preparation before surgery (the table 2).
84.5 % of patients in the main group showed longer hospital stay before surgery (> 3 days) in comparison with 74.5 % in the comparison group (p = 0.04) (the table 2). It was determined by preliminary correction of somatic pathology in the profile unit.
The following complications of intramedullary fixation were identified in the postsurgical period: thrombophlebitis and thrombosis of deep veins of lower extremities were identified in 3 and 10 cases, with 2 cases with embolism danger (the table 2). It required for vascular surgical interventions. After osteosynthesis, one case in each group was associated with metal construct migration, and migration of one cannulated screw among three implanted ones. With consideration of estimation of intrasurgical compression, the patients received recurrent installment of the screw. Contact dermatitis was registered in 1 case in the comparison group.
After hip replacement, thrombophlebitis and thrombosis of lower extremity veins were registered in 4 and 7 cases (with 1 case of embolism-threatening condition, which required for vascular surgical intervention) in the main and comparison groups correspondingly.
The high incidence of thrombophlebitis and thrombosis was determined by total control with use of duplex scanning, which was included into the postsurgical management algorithm for patients with proximal femoral fractures.
In the comparison group, one patient had a dislocation of head of the hip joint endoprosthesis. Closed reduction of the endoprosthesis dislocation was conducted after additional X-ray examination.
A local infection in the surgical wound was registered in one case in the comparison group. Contact dermatitis was in one case. The patients received the conservative treatment. Its results were positive.
The total number of postsurgical complications in the patients with proximal femoral fractures was 8.8 % and 15.2 % in the main and comparison groups correspondingly (the table 2). Therefore, the patients of the main group showed 1.7 times less registered postsurgical complications (p = 0.009) as compared to the comparison group.
There were not any statistically significant differences in the values of ICU stay of patients with proximal femoral fractures after surgery in the compared groups (the table 2).
In the study group, the hospital stay was 1.3 times shorter (p = 0.04) in relation to this value in the comparison group. One should note that duration of the postsurgical period in the groups allowed partial removal of suture and teaching the patients to walk with crutches.
All patients (100 %) were discharged for outpatient treatment. Their condition was satisfactory, without complications and lethal outcomes at the moment of discharge.
Clinical estimation of functional outcomes of surgical management of proximal femoral fractures was conducted one year after surgery in 84 (93 %) patients in the main group and in 135 available (93 %) patients in the comparison group. The functional outcomes of osteosynthesis and hip replacement were estimated in 35 % and 57 % of patients with proximal femoral fractures in each group correspondingly.
The comparative analysis of functional results of treatment of proximal femoral fractures after Goodwin osteosynthesis showed the most favorable outcomes of function recovery in the main group. It was manifested by increasing number of fine and good results – 2.8 times (χ2 = 12.8, p = 0.0025) and 2 times (χ2 = 3.8, p = 0.05), whereas the comparison group showed the highest satisfactory result (the table 3).
Table 3
Estimation of functional results of treatment of patients with proximal femoral fractures 1 year after osteosynthesis (R.A. Goodwin, 1968) [19]
|
Estimation of results |
Main group n = 32 |
Comparison group * n = 51 |
χ2 (p) |
||
|
n |
% |
n |
% |
|
|
|
Fine |
14 |
43.8 |
8 |
15.7 * |
12.8 (0.025) |
|
Good |
11 |
34.4 |
9 |
17.6 * |
3.8 (0.05) |
|
Satisfactory |
5 |
15.6 |
29 |
56.9 * |
30.1 (0.05) |
|
Poor |
1 |
3.1 |
4 |
7.8 |
0.02 (0.89) |
|
Unsatisfactory |
1 |
3.1 |
1 |
2.0 |
0.00016 (0.89) |
|
Total |
32 |
100 |
51 |
100 |
|
Note: * - comparison of groups with χ2, p < 0.05
The highest quantitative values of improvement in function of the lower extremity and the hip joint (Harris score) were identified in the main group of patients with proximal femoral fractures who received hip joint replacement. The results of treatment of 49 patients of the main group were considered as fine and good (the table 4). Harris score showed more fine and good results – 1.5 times (χ2 = 12.6, p = 0.05) and 1.2 times (χ2 = 6.0, p = 0.05) higher correspondingly than in the comparison group. One should note that the best functional result was received in patients of the main group by means of decreasing potential and actual risk of complications. It allowed active rehabilitation in early postsurgical period. Higher rate of fine and good postsurgical values of Harris score in the main group shows earlier and more full recovery of function of the injured lower extremity and the joint, resulting in faster return to daily life.
Table 4
Estimation of functional results of treatment of patients with proximal femoral fractures 1 year after hip surgery (W.H.Harris, 1969) [20]
|
Estimation of results |
Main group n = 52 |
Comparison group * n = 84 |
χ2 (p) |
||||
|
Estimation |
Points |
n |
% |
n |
% |
|
|
|
Fine |
90-100 |
19 |
36.5 |
20 |
23.8* |
12.6 (0.05) |
|
|
Good |
80-89 |
30 |
58.0 |
40 |
47.6* |
6.0 (0.05) |
|
|
Satisfactory |
70-79 |
3 |
5.5 |
24 |
28.6* |
31.5 (0.025) |
|
|
Unsatisfactory |
ìåíåå 70 less than 70 |
- |
- |
- |
- |
|
|
|
Total |
100 |
52 |
100 |
84 |
100 |
|
|
Note: * – comparison of groups with χ2, p < 0.05.
DISCUSSION
The study confirmed the possibility of clinical administration of the prediction model of the risk of postsurgical complications in patients with proximal femoral fractures. Our model accurately predicts the risk of postsurgical complications for each patient with consideration of such parameters as age, gender, comorbid status, ASA severity class, type of proximal femoral fracture (the table 1). The study showed that the estimation system in view of the nomogram [14] allows efficient differentiation of subgroups into of different risk (low, average and high) and fast and easy identification of high risk patients during hospital stay.
One should note that our efforts for development of the model of prediction of complications in patients with injuries are not unprecedented. For the last 30 years, Trauma and Injury Severity Score (TRISS) [23] was the main technique for estimation of trauma outcome. TRISS model, which was developed with logistic regression, was developed for prediction of probability of post-injury survival with consideration of age, ISS [17] and Revised Trauma Score (RTS). Whereas TRISS was widely used for estimation and comparison of trauma outcomes, it was criticized for important disadvantages in traumatology literature. For the last 20 years, it was highly criticized for that fact that it is based on the data basis of 80s, on Major Trauma Outcomes Study. Since the trauma management system has significantly changed during this period, the modern predictive significance of TRISS was called into question. It was partially associated with development of the new indices in 1990 and in 1995, but necessity for continuous and multiple updates limited their practical administration [26]. Another disadvantage of TRISS is its inability to predict the risk of postsurgical complications. A Severity Characterization of Trauma (ASCOT) [27] was acknowledged as the improved predictor of complications and outcomes, but the complexity of calculations limits its wide use. Other attempts of prediction of risk of post-injury complications were made [28], but, unfortunately, there is not easy and statistically accurate modern model of potential risk of complications after femoral fractures. The advantage of use of our prediction model of risk of postsurgical complications in patients with proximal femoral fractures with consideration of age, gender, comorbid status and condition severity consists in receive of more objective information on a patient before surgical treatment and correct optimization of surgical treatment with minimal risks.
One should acknowledge the disadvantages of our prediction model. Possibly, the main disadvantage is the fact that model was developed and tested in relation to index (inhospital) admission. The period of follow-up of clinical and functional results within one year after surgery is insufficient for the fullest estimation of treatment outcomes. There is an unexplored fact: is it possible to use the prediction model for estimation of intermediate and remote results of surgical treatment of proximal femoral fractures at the outpatient stage? The model is only the predictor of probability of complications, and it does not estimate the return to previous functional status and does not predict it and quality of patient’s life in the future. But this disadvantage only means that this information has to be tested in further prospective studies with collection and analysis of data on outcomes of femoral bone fractures that can be important for making a decision of preventive and rehabilitation measures.
CONCLUSION
The results of the study show higher efficiency of surgical treatment (fixation or primary total hip replacement) in patients with proximal femoral fractures with consideration of potential estimation of risk of postsurgical complications.
The clinical use of the prediction model of risk of postsurgical complications for efficient surgical treatment (osteosynthesis or primary total replacement) of patients with proximal femoral fractures was accompanied by 1.7-fold decrease (p = 0.009) in complications, 1.3- fold decrease (p = 0.04) in time of treatment, the increase in fine and good functional results 2 times (p = 0.001) and 1.4 times (p = 0.05) correspondingly.
Our prediction model of the risk of postsurgical complications in patients with proximal femoral fractures with use of combination of parameters (age, gender, a comorbidity category, ASA severity class, a fracture type) allows precise prediction of development of possible complications in the postsurgical period. Moreover, the estimation system with the nomogram allows efficient differentiation of patients into subgroups of different risk (low, average, high), and fast and easy identification of patients with high risk during admission period. Owing to continuous prediction of a probability of complications with use of small range of parameters, it can be used as a tool of dynamic observation of patients’ condition.
Our surgical management algorithm (osteosynthesis or primary total hip replacement) for patients with proximal femoral fractures with consideration of predictive risk of postsurgical complications is a simple and easy for understanding and daily practical use, with high potential for development of guidelines for treatment of proximal femoral fractures.
Information on financing and conflict of interests
The study was conducted without sponsorship.
The authors declare the absence of any clear or potential conflicts of interests relating to publication of this article.