THE PREDICTION MODEL OF POTENTIAL RISK OF COMPLICATIONS IN PATIENTS WITH PROXIMAL FEMORAL FRACTURE Agadzhanyan V.V., Milyukov A.Yu., Ustyantsev D.D., Gilev Ya.Kh.
Regional Clinical Center of Miners’ Health Protection, Leninsk-Kuznetsky, Russia,
Tsyvyan Novosibirsk Research Institute of Traumatology and Orthopedics, Novosibirsk, Russia
Currently, the incidence of musculoskeletal system diseases is increasing, the high rate of injuries is persistent, and their social consequences are high (temporary working incapability and disability) [1, 2].
According to the assumptions, the amount of persons older 65 will increase twice by 2025 [3]. Older persons suffer from degenerative and dystrophic diseases such as osteoarthrosis, osteochondrosis and osteoporosis with 60 % higher rate than in young persons [4]. With ageing of population, the number of hip joint fractures increases [5, 6]. The rate of femoral neck fractures is increasing: 75.3-80.2 % against 19.8-24.7 % of acetabular fractures [4].
Although proximal femoral fractures consist less than 20 % of all osteoporotic fractures, they lead to death in most cases with fractures in persons older 50 [7]. In patients with hip joint fractures, the risk of death is 20-30 % within one year, whereas 30-day mortality is 5-10 % [8].
Most authors note the high mortality within one year after trauma – 63 % for conservative therapy in older patients with proximal femoral fractures [9]. Exacerbation and decompensation of patients’ condition owing to concurrent diseases were the causes of death determined by increasing cardiovascular insufficiency, development of cerebral perfusion disorders, pneumonia and bedsores [10].
According to predictions by L.J. Melton, the amount of proximal femoral fractures will be about 6-6.5 million by 2050 [11].
According to recent systematic reviews and metaanalysis, some predictors are associated with death after surgery for hip joint fractures, including an injury type, older age, male sex, presurgical mobility, cognitive worsening and presence of concurrent diseases [5, 12, 13, 14].
Considering some reasons, proximal femoral fractures require for specific approach to treatment since the injury in the older patient is a complex surgical, therapeutic, social and psychological problem, which is to be solved by physicians of different specialties and social workers [1, 10].
The data shows that the higher percentage of patients with proximal femoral fractures receives the treatment in outpatient conditions [10]. It is well known that the fracture union is almost impossible without osteosynthesis [15].
It is known that the possibility of high risk of postsurgical complications in patients with proximal femoral fractures is directly proportional to the increase in age and comorbidity index [16]. The nomogram for clinical estimation of the risk of complications in patients with proximal femoral injuries and concurrent diseases can be used in combination with screening of functional and physiological parameters [16].
The next stage was the development of a method for choice of surgical management for proximal femoral fractures on the basis of estimation of the potential risk of complications according to the nomogram and estimation of severity class according to American Society of Anaesthesiologists (ASA).
The study objective was to develop the model for selecting the surgical management (osteosynthesis or primary total arthroplasty) for proximal femoral fractures and identification of patients with high risk of postsurgical complications with consideration of age, gender and comorbid status in combination with screening of previous functional and physiological parameters and estimation of clinical results of use of this model.
MATERIALS AND METHODS
The retrospective study included the results of the complex examination and surgical treatment of 161 patients with proximal femoral fractures who were observed by us in 2013-2016. The data was extracted from the electronic database of the Medical Information System (MIS) of Regional Clinical Center of Miners’ Health Protection.
The inclusion criteria for the patients with proximal femoral fractures were the age > 18, ISS (Injury Severity Score) ≤ 15 [17], absence of transfers to other clinics, more than 24 hours of stay in the clinic. The general mortality in the included patients was not registered.
In concordance with ICD-10, proximal femoral fractures were verified on the basis of complaints, the data of physical examination and radiography of the hip in two standard planes. A type of proximal femoral fracture was determined with the modified classification by Kaplan A.V. (1967) [18].
The surgical management included the osteosynthesis with intramedullary nails PFN, the cannulated screws and primary total replacement of the hip joint in concordance with the modern guidelines [19].
The severity class of the patient was estimated with the classification of the objective status from ASA [20].
Parameters
The MIS database included all variables used in the study, individually for each patient with a proximal femoral fracture.
The stratifying variables were: the age (18-64 – young age; 65 and older – older age), gender, comorbidity status estimated with 3 categories (0 – absence of concurrent diseases, mean chronical condition (1-2 concurrent diseases), multiple morbidity (3+ concurrent diseases), a fracture type, a class of the patients’ severity condition with ASA [20], a surgery type, hospital stay after surgery.
The extreme values of the vital parameters and the laboratory results from MIS database were used for estimation of ASA condition severity. The variables, which were absent in 10 % of the patients, were excluded from the further analysis. All variables were analyzed as dichotomic values for simplification of inclusion into the final prediction model.
The objective estimation of adequate preparation of the patients to surgery was conducted with the nomogram developed by us [16]. The predictive risk of complications was estimated in relation to increasing age and the comorbidity categories for men and women.
The study corresponded to World Medical Association Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects, 2013, and to the Rues for Clinical Practice in the Russian Federation (the Order by Russian Heath Ministry, June 19, 2003, No.266), and was approved by the local ethical committee. Since the study was of surveillance type, the informed consent was not required.
Statistical analysis
The statistical analysis of the data was conducted with IBM SPSS Statistics 21 (Statistical Product and Service Solutions – SPSS).
During statistical analysis, the extensive coefficients (%) characterizing the relation of the parts to the whole were estimated. The qualitative signs were presented as absolute and relative (%) values. The qualitative variables were presented as the mean arithmetic (M) and quadratic deviation of the mean arithmetic (SD) in the ordered sample, in view of Me (LQ-UQ), where Me – the median, (LQ-UQ) – interquartile range (LQ – 25%, UQ – 75% quartiles). The critical level of significance (p) was < 0.05 for testing the statistical hypotheses.
The randomized heterogeneous combined sample was used as a method for formation of the sampled population. The relationships were identified with univariate and multivariate logistical regression.
The primary result for inclusion of the variables into the univariate logistical regression analysis was the availability of estimation of the risk of postsurgical complications in the patients with proximal femoral fractures.
The univariate logistical regression was used for analysis of each variable (age, gender, a comorbidity category, a fracture type, ASA class severity, surgery type) and the rate of development of complications which was used as the borderline value for inclusion of a variable into the multivariate model.
Multiple logistical regression was used for identification of the risk factors of postsurgical complications. The variables were fixed in the final predictive multivariate model, if p < 0.05.
The point rating system was developed on the basis of the agreement coefficients of the multivariate analysis to simplify the possibility of use in clinical practice. The value of 0.5 unit (point) was assigned to the variables, when the coefficient was less than 0.75; 1 unit point – if the coefficient was between 0.75 and 1.25; 1.5 unit (point) – if the coefficient was from 1.25 to 1.75, and 2 units (points) – for the coefficient > 1.75.
The final model for prediction of the risk of postsurgical complications after proximal femoral fractures was the derivative from the regression formula, which was created on the basis of the points in combination with the values of agreement coefficients of the multivariate analysis. The sample median was used for using the model instead of the missed values of the risk.
The discriminating ability of the model was estimated with ROC-curve. The area under ROC depends on the sensitivity and the specificity. It changes within 0.50-1.00. This is a criterion of efficiency of selecting the patients who correspond to a selected criterion (the potential risk of complications after proximal femoral fractures).
ROC 0.50 predicts the probable chance, whereas ROC 1.00 – the value of absolute recognition. ROC-curve within 0.70-0.79 presents the admissible recognition in the model of prediction of complications (within 0.80-0.90 – excellent).
The total estimation of the model and the real data was conducted with Hosmer and Lemeshow goodness-of-fit test, which compares the observed (actual) and expected (potential) values of the risk of complications according to the prediction deciles. The final values (p < 0.05) testified that the difference between the observed risk and the predicted risk of complications was higher than it was expected, and it showed the disadvantage in compliance in the model.
The patients with proximal femoral fractures were distributed into the groups of low (< 10 %), middle (10-30 %) and high (> 30 %) risk in concordance with the actual values of the complications registered by us. Henceforth, the observed (actual) numbers of the complications were compared with the risks of possible complications which were calculated in our model.
RESULTS AND DISCUSSION
Results
From January 2013 to December 2016, 161 patients with proximal femoral fractures were operated in compliance with the criteria (ISS ≤ 15, age ≥ 18, absence of transfer to other clinics, stay in the clinic ≥ 24 hours). The general lethality was not registered in the included patients.
In case of absence of any data of the outpatient status at the moment of hospital admission, for example, body mass, serum albumin and others (approximately in 10 %), the patients were excluded from the study. After exclusion of insufficient medical records (n = 16 (9.9 %)), the final analysis included the medical records of 145 patients with proximal femoral injuries.
The study group included 30 (20.7 %) young patients (age of 18-64), 115 (79.3 %) older patients older patients (≥ 65), most patients of male sex (84 patients, 57.9 %).
The types of proximal femoral fractures were as described below: medial fractures or intraarticular femoral neck fractures – 111 patients (76.6 %) (subcapital fracture – 32 (28.8 %), transcervical – 47 (42.3 %), basicervical – 32 (28.9 %)); lateral or trochanteric extraarticular fractures – 34 patients (23.4 %) (transtrochanteric – 24 (70.6 %), subtrochanteric – 10 (29.4 %)).
The young patients (age of 18-64) with medial and lateral fractures of the proximal femoral bone received 6 procedures of primary total hip replacement, 10 patients received the femoral bone fixation with the cannulated screws, 15 patients – the femoral bone fixation with PFN (18 %, 36 % and 46 % of all surgeries in this group).
The older patients (≥ 65) received the hip joint replacement in 84 cases (74 %).
Most patients were operated two days after admission (n = 135; 93.1 %).
The estimation of the comorbidity categories showed the absence of concurrent diseases in 10 (6.9 %) patients (0 – a comorbidity category), 80 patients (55.2 %) had the comorbidity category 1-2, 55 patients (37.9 %) had the concurrent pathology with three and more diseases (3+).
In concordance with the ASA objective status classification, the severity categories 1, 2, 3 and 4 were registered in 13 (8.9 %), 24 (16.6 %), 70 (48.5 %) and 38 (26.2 %) patients with proximal femoral fractures correspondingly.
The mean duration of hospital stay was 14.5 (17.6) days.
The table 1 shows the characteristics of the results of the univariate logistical regression analysis for determination of potential factors relating to the risk of complications after surgical management of the patients with proximal femoral fractures.
Table 1
Characteristics of results of univariate logistic regression analysis for determination of potential risk factors of complications after surgical treatment of patients with proximal femoral fractures (n = 145)
|
Parameters
|
n (%) 1 |
Amount of complications, abs. (%)1 |
OR (95% CI) |
p |
|
Age: young age (18-64)
older age (65+)
|
30 (20.7)
115 (79.3) |
4 (2.8)
18 (12.4) |
2.23 (1.32 – 3.76) |
0.007 |
|
Gender: |
|
|
|
|
|
men women
|
84 (57.9) 61 (42.1) |
15 (10.3) 7 (4.8) |
1.74 (1.03 - 2.92) |
0.037 |
|
|
|
|
|
|
|
Fracture type: |
|
|
|
|
|
medial |
111 (76.6) |
14 (9.7) |
1.59 (0.93 – 2.72) |
0.090 |
|
lateral |
34 (23.4) |
8 (5.5) |
|
|
|
|
|
|
|
|
|
Comorbidity (concurrent diseases before injury)1: |
|
|
|
|
|
no concurrent diseases (0) |
10 (6.9) |
1 (0.7) |
|
|
|
1-2 concurrent diseases |
80 (55.2) |
5 (3.4) |
3.11 (1.67 - 5.78) |
< 0.001 |
|
3+ concurrent diseases |
55 (37.9) |
16 (11.0) |
3.73 (1.99 - 6.98) |
< 0.001 |
|
|
||||
|
|
|
|
|
|
|
|
|
|
|
|
|
ASA condition severity class* 1 |
13 (9) |
1 (0.7) |
|
|
|
2 |
24 (16.6) |
3 (2.1) |
2.25 (1.09 - 4.65) |
< 0.027 |
|
3 |
70 (4.8) |
12 (8.3) |
3.24 (1.05 - 10.2) |
< 0.041 |
|
4 |
38 (26.2) |
6 (4.1) |
4.27 (1.91 - 9.56) |
< 0.001 |
|
|
|
|
|
|
|
Surgery type |
|
|
|
|
|
osteosynthesis |
55 (37.9) |
12 (8.3) |
|
|
|
prosthetics |
90 (62.1) |
10 (6.9) |
1.43 (1.09 – 2.84) |
0.071 |
|
Days before surgery |
|
|
|
|
|
1 |
10 (6.9) |
2 (1.4) |
|
|
|
2 |
27 (18.6) |
5 (3.4) |
1.32 (0.60 – 2.89) |
0.487 |
|
3+ |
108 (74.5) |
15 (10.3) |
1.46 (0.87 – 2.45) |
0.150 |
Note:
1Percentage ratio is has been calculated with consideration of all patients.
*Condition severity class according to objective status of American Society of Anesthesiologists (ASA) [20].
p < 0.05 values are separated with bold type; CI – confidence interval; OR – odds ratio.
The univariate logistical regression analysis identified some values, which supposed the higher risk of postsurgical complications in the patients with proximal femoral fractures. The positive, most significant relationship was found: the age > 65 (p = 0.007), gender – male (p = 0.037), the comorbidity categories: 1-2 concurrent diseases (p < 0.001) and 3+ concurrent diseases (p < 0.001), ASA severity class 2 (p = 0.027), ASA severity classes 3 and 4 (p < 0.001).
A fracture type, a surgery type and time of surgery did not exert significant influence on the risk of postsurgical complications (ð = 0.09, ð = 0.071, ð = 0.082).
The final values of our model for prediction of the risk of postsurgical complications in the patients with proximal femoral fractures are presented in the table 2. Six parameters were selected as the predictors of postsurgical complications in patients with proximal femoral fractures: gender, age, a comorbidity category, potential risk of complications, ASA severity class, a fracture type.
Table 2
Results of multiple logistic regression of risk factors of complications after surgical treatment of patients with proximal femoral fractures (n = 145)
|
Values |
Coefficient |
Points1 |
OR (95% CI) |
p |
|
Age: young age(18-64 years) older age(65+ year) |
0.4467 1.8240 |
0.5 2 |
1.72 (1.02 – 2.89) 6.20 (1.51 – 33.39) |
0.037 < 0.001 |
|
Gender: |
|
|
|
|
|
men women |
1.8505 0.6524 |
2 1 |
2.34 (1.27 – 4.33) 1.92 (1.04 – 3.54) |
< 0.001 0.037 |
|
|
|
|
|
|
|
Fracture type: |
|
|
|
|
|
medial |
1.4526 |
1.5 |
1.27 (1.13 - 3.56) |
0.041 |
|
lateral |
1.1050 |
1 |
1.04 (0.86 – 2.67) |
0.034 |
|
Comorbidity (concurrent diseases before injury)1: |
|
|
|
|
|
no concurrent diseases (0) |
0.4534 |
0.5 |
0.72 (0.43 – 1.860) |
0.027 |
|
1-2 concurrent diseases |
1.9985 |
2 |
7.38 (2.46 -22.14) |
< 0.001 |
|
3+ concurrent diseases |
2.0727 |
2 |
7.95 (2.66 -23.74) |
< 0.001 |
|
ASA condition severity class* |
|
|
|
|
|
1 |
0.3279 |
0.5 |
1.03 (0.47 – 3.4) |
< 0.027 |
|
2 |
0.8305 |
1 |
1.32 (0.6 – 2.89) |
< 0.001 |
|
3 |
1.9346 |
2 |
6.87 (1.8 – 27.39) |
< 0.001 |
|
4 |
2.0335 |
2 |
7.75 (2.48 – 25.3) |
< 0.001 |
Note: 1 Scoring estimation was developed on the basis of agreement coefficients of multivariate analysis with aim of simplification of possible clinical use with Hosmer and Lemeshow Goodness of Fit Test.
The value of 0.5 unit (points) was prescribed for the variables, when the coefficient was less than 0.75; 1 unit (point) – if the coefficient was between 0.75 and 1.25; 1.5 unit (point) – for the coefficient from 1.25 to 1.75 and 2 units (points) for the coefficient > 1.
*Severity class of patient’s condition according to classification of objective status of American Society of Anesthesiologists (ASA) [20].
CI – confidence interval, OR – odds ratio.
Among our patients, older age (65+), male sex, presence of 3+ diseases, and ASA severity classes 3 and 4 showed the highest agreement coefficients, as shown in the table 2.
We found the maximal values between the contribution coefficients (0.5-2 units) since they could differentiate the risk of postsurgical complications lower and higher 30 % (the table 2).
Using the simple scoring system (1-4 points), with ASA severity classes 1, 2, 3 and 4, we made the correlation to the agreement coefficient values (0.5; 1; 1.5; 2 units corresponded to 1; 2; 3; 4 points) (the table 2). On the basis of the values and the ratings of the agreement coefficients, the items for each parameter were located as it shown in the table 2. The scoring system did not influence on the predicted outcome.
The area under ROC-curve was 0.81 for the multivariate analysis. Hosmer-Lemeshow test was not significant (p = 0.343).
Using the simple scoring system (1-4 points), we separated the groups of low, middle and high risk of postsurgical complications (the table 3). The patients with low risk had approximately 1 point and the maximal predicted risk of postsurgical complications in 2.1-8.9 %; the patients with middle risk – 2-3 points, and the expected risk after surgery – 13.9-30.3 %; the patients with high risk (n = 55) (> 3 points) showed 34.7 % and higher predicted risk of complication (the table 3).
Table 3
Characteristics of risk groups (low, middle and high) in concordance with potential and observed (factual) values of complications after surgical treatment of patients with proximal femoral fractures (n = 145)
|
Points 1 |
Potential risk of complications 2, (%) |
Risk groups
|
Observed (factual) complications 3, (%) |
ASA severity class 4 |
|
0 0.5 1 1.5
|
2.1 3.5 5.6 8.9 |
Low (0-10 %)
|
2.8 |
1 |
|
2 2.5 3.0 |
13.9 21.0 30.3 |
Middle (11-30 %) |
12.1 |
2 3 3 |
|
3.5 4.0
|
34.7 39.6
|
High (> 30 %) |
33.4 |
4 4 |
Note:
1Scoring estimation developed on the basis agreement coefficients of multivariate analysis (table 2).
2Potential risk of complications has been calculated with consideration of all patients with nomogram [16].
3Percentage ratio of complications has been calculated with consideration of all patients.
4Severity class of patient’s condition according to classification of objective status of American Society of Anesthesiologists (ASA)
After distribution into the risk groups, 10 patients (6.9 %) were in the low risk group, where the risk of postsurgical complications was 2.8 %. The middle risk group included 80 patients (55.2 %), with the risk of complications of 12.1 %.The high risk group showed the risk of postsurgical complications of 33.4 % and included 55 patients (37.9 %).
The scoring system did not influence on the predicted result in comparison with the logarithmic distribution model (ROC 0.806 in comparison with 0.813 in the derived set; Hosmer-Lemeshow p > 0.05).
Therefore, the high risk of postsurgical complications in the patients with proximal femoral fractures is closely related to the increase in age, comorbidity index, ASA severity class, gender and fracture types.
The results of the conducted study were the basis for development of a method for selection of surgical techniques for proximal femoral fractures in patients with concurrent diseases on the basis of calculation of the potential risk of complications in concordance with our nomogram [16] with consideration of age, gender, comorbidity categories, estimation of severity using ASA classification [20].
Choice of surgical management for proximal femoral fractures [21]
The offered method is used as described below. At admission, the anamnesis data is collected with estimation of concurrent diseases, age; the patient is examined, and surgical techniques are selected. Additionally, gender is considered. The young age is 18-64 years, 65 and more – older age. A comorbidity category is determined: no concurrent diseases – healthy, 1-2 concurrent diseases – average chronic condition, 3 and more concurrent diseases – multiple morbidity.
The potential risk of complications is estimated with the nomogram (Fig. 1). The nomogram is presented separately for the women (Fig. 1a) and the men (Fig. 1b). Dimension of Y-axis corresponds to the potential risk of complications with division value of 10 % and the graduation scale from 0 to 60 %. Dimension of X-axis corresponds to age, with division value of 10 years and the graduation scale from 0 to 90 years. The third axis is parallel to Y-axis. It is located in the end of X-axis and corresponds to ASA severity class, with division value of one class and the scale from 0 to 6th class. The nomogram field is divided into two zones with a line going in parallel with X-axis at the level of 30 % in Y-axis. The nomogram presents the charts corresponding to three categories of comorbidity: 0 – no concurrent diseases, 1-2 concurrent diseases, 3 and more concurrent diseases. The horizontal direct line (risk of 30 %) shows the possibilities of using the nomogram for calculation of the potential risk of complications. So, all values, which are higher the values from the direct line of 30 % have the high risks of complications with consideration of gender and age of trauma patients. For example, for men at the age of 50 (Fig, 1b), this value is found in X-axis; the vertical line is drawn to the intersecting point with the chart (3 concurrent diseases). From this point, a horizontal line is drawn to Y-axis. The risk of complications is 30 %; similarly, for the age of 55, with two concurrent diseases at least, and for all men at the age of 70 and older. Also the risk of complications is estimated with consideration of age and concurrent pathology in women in other nomogram considering the gender (Fig. 1a).
The clinical, laboratory and instrumental study is conducted. Additionally, the severity of patient’s condition is estimated with ASA classification (the class 1: a normal healthy patient; the class 2: a patient with a mild systemic disease; the class 3: a patient with severe systemic disease, which is not dangerous for life; the class 4: a patient with a severe systemic disease, which is life threatening; the class 5: a dying patient requiring for urgent surgical management for vital indications; the class 6: brain death) [20]. A fracture type was estimated with the modified classification by A.V. Kaplan [18].
The surgical techniques are selected using the combination of six parameters: gender, age, a comorbidity category, potential risk of complications, ASA class, a fracture type, with use of the nomogram for the women (Fig. 1a) and for the men (Fig. 1b).
Figure 1
The choice of surgical management for proximal femoral fractures with use of the nomogram of potential risk of complications in women (a) and men (b):
a) The nomogram with consideration of age, concurrent diseases and ASA severity class in women;
b) The nomogram with consideration of age, gender, concurrent diseases and ASA severity class in men.
The nomogram presents the charts corresponding to three comorbidity categories: 0 – absence of concurrent diseases, 1-2 concurrent diseases, 3 and more concurrent diseases. The field of the nomogram is divided into two zones with the horizontal line, which goes in parallel to the abscissa axis (X) at level of the ordinate axis (Y) and corresponds to 30 % of potential risk of complications. The third axis is parallel to the ordinate axis and is located in the end of the abscissa axis and corresponds to ASA severity class [20] with division value of one class and the scale from 0 to 6 class.
Note: ASA severity class: class 1 – normal healthy patient; class 2 –
patient with mild systemic disease; class 3 – patient with severe systemic
disease, which is not life-threatening; class 4 – patient with severe systemic
disease, which is life-threatening; class 5 – dying patient who needs for acute
surgical management according to urgent indications; class 6 - brain death
registered.
If the combination of the ASA severity class 4 with the potential risk of complications is higher than 30 %, the surgical interventions are not performed. For ASA class 3 with potential risk of complications < 30 % in medial and lateral fractures, low invasive osteosynthesis for the proximal femoral bone is conducted. The choice of techniques is determined by characteristics of a fracture (screws or PFN).
With ASA severity classes 1-2 with potential risk of complications < 30 %, a patient with lateral and basicervical fractures receives the osteosynthesis, for medial fractures (subcapital and transcervical) – total hip replacement.
The practical use of the method is confirmed by the clinical cases.
Clinical case 1
The patient M., year of birth 1948, was admitted to the traumatology and orthopedics unit No.2, Regional Clinical Center of Miners’ Health Protection, on September 20, 2015. The diagnosis was: “A closed basicervical fracture of the left femoral neck (Fig. 2a). The period of the injury – 3 days. The patient did not seek medical advice.
After admission, the history was taken: osteoporosis. Chronic gastritis, the remission stage. Gastric ulcer in the remission phase. Hypertonic disease of the degree 2, the second stage, risk 3. Postthrombophlebitic syndrome of the right lower extremity, chronic venous insufficiency of degree 2. Cardiac insufficiency of functional class 2.
The age and the gender were determined: the man, age of 68, older age. A comorbidity category: multiple morbidity (3+ concurrent diseases). The clinical examination was conducted. The nomogram showed 38 % risk of complications (Fig. 1b). The estimation showed ASA class 4. Therefore, the risks of surgical intervention relating to the age and concurrent pathology were high and non-compensated.
Owing to high risks of a surgical intervention, the patient received the conservative symptomatic treatment of cardiovascular disease controlled by the cardiologist and the therapeutists during 7 days. After correction of somatic pathology, the recurrent estimation of the risk of complications was conducted (24 %, ASA 3 class, moderate risk). The planned surgery was carried out: closed reposition of the basicervical fracture of the left femoral bone, low invasive osteosynthesis with the cannulated screws and electronic optical amplification (Fig. 2b). The perisurgical complications were absent. The movements were within the full range (Fig. 2c). The patient could move next day after the surgery. The hospital period was reduced to 6 days. The patient was discharged in good condition.
Figure 2
Clinical case 1:
a) X-ray image of the patient M. with basilar fracture of left femoral neck;
b) X-ray image of the patient M. after low-invasive osteosynthesis with cannulated screws;
c) a picture of the patient M. two years after treatment.
| Figure2a X-ray image of the patient M. with basilar fracture of left femoral neck | Figure 2b X-ray image of the patient M. after low-invasive osteosynthesis with cannulated screws | Figure 2c a picture of the patient M. two years after treatment |
 |  |  |
Clinical case 2
The patient K., age of 39, was admitted to the traumatology and orthopedics unit No.2, Regional Clinical Center of Miners’ Health Protection. The diagnosis was: “A closed transtrochanteric fracture of the left femoral bone” (Fig. 3a).
The anamnesis data was collected after admission, and some concurrent diseases were identified. The secondary diagnosis: chronic gastritis in remission phase. Diabetes mellitus, type 2, compensated.
The age and the gender were determined: the man, age of 39, young age. A comorbidity category: 1-2 concurrent diseases. The clinical examination was conducted. The nomogram showed 18 % risk of complications (Fig. 1b). The severity degree was ASA 2. Therefore, the risks of a surgical intervention relating to the age and the concurrent pathology were low. On the second day, the patient received: closed reposition of a transtrochanteric fracture of the left femoral bone, intramedullary fixation of the left femoral bone with the proximal locked nail under electron-optical image intensifier (Fig. 3b). There were not any postsurgical complications, the movements were within the full range (Fig. 3c).The patient could move on the next day after the surgery. The hospital stay was reduced to 8 days. The condition was good at the moment of discharge.
Figure 3
Clinical case 2
a) X-ray image of the patient K. with a closed pertrochanteric fracture of the left femur;
b) X-ray image of the patient K. after intramedullary fixation of the left femur with proximal locked nail;
c) a picture of the patient K. 1 year and 9 months after treatment.
| Figure 3a X-ray image of the patient K. with a closed pertrochanteric fracture of the left femur | Figure 3b X-ray image of the patient K. after intramedullary fixation of the left femur with proximal locked nail | Figure 3c a picture of the patient K. 1 year and 9 months after treatment |
 |  |  |
Clinical case 3
The patient V., age of 74, was admitted to the traumatology and orthopedics unit No.2, Regional Clinical Center of Miners’ Health Protection. The diagnosis was: “A closed subcapital fracture of the left femoral neck” (Fig. 4a).
The anamnesis data was collected after admission. Some concurrent diseases were identified: coronary heart disease. Postinfarction cardiosclerosis since 2007. CHF, functional class 2. Hypertonic disease 3, risk 4. Tachysystolia, normophorm. Cerebrovascular disease, degree 1. Postthrombophlebitic syndrome to the left. Obesity of degree 2.
The gender and the age were determined: the woman, age of 74, older age. A comorbidity category: multiple morbidity (3+ concurrent diseases). The clinical examination was conducted. The nomogram showed 37 % risk of complications (Fig. 1a). ASA 4 class severity. Therefore, the risks of surgical intervention relating the age and concurrent pathology are still high and non-compensated. Owing to the high risk of a surgical intervention, the patient received the conservative and symptomatic treatment of the cardiovascular diseases and was observed by the cardiologist, the neurologist and the therapeutists during 9 days. After correction of somatic pathology, the risk of complications was estimated again – 24 %, ASA class 2, i.e. moderate risks. Considering the characteristics of the injury, the patient received the planned total replacement of the left hip joint (Fig. 4b). There were not any postsurgical complications. The patient could move on the day 3 after the surgery. The hospital stay was 19 days. She was discharged in good condition.
Figure 4
Clinical case 3:
a) X-ray image of the patient V. with a closed subcapital fracture of the left femur neck;
b) X-ray of the patient V. after total
replacement of the left hip joint.
| Figure 4a X-ray image of the patient V. with a closed subcapital fracture of the left femur neck | Figure 4b X-ray of the patient V. after total replacement of the left hip joint |
 |  |
DISCUSSION
Along with the increase in the incidence of proximal femoral fractures [10, 11], the necessity for estimation of the predictor risk of postsurgical complications increases.
Multiple studies are dedicated to the risk factors of postsurgical complications and mortality in patients with proximal femoral fractures [12, 14, 15]. However it is difficult to estimate the risk of each individual patient. Such model for risk prediction is required (better at the moment of admission) that is used by the attending physician. Moreover, the model for prediction of postsurgical complications in patients with proximal femoral fractures can be used as a testing method in comparison of outcomes in various medical facilities.
Many authors confirmed the importance of the problem and showed the role of appropriate therapeutic examination and treatment for improvement in life quality and for further prediction after surgery [8, 14, 15]. Few studies are related to selection of optimal surgical techniques for proximal femoral fractures. Some authors offered selecting the surgical management of the femoral neck with consideration of age and general condition, but they did not offer to systematize the risks of concurrent diseases and a surgical intervention for determination of subsequent surgical management [22, 23].
As we know, there are several models for risk prediction in surgical treatment of hip fractures [24-26]. Orthopaedic Physiologic and Operative Severity Score for the Enumeration of Mortality and Morbidity (O-POSSUM) uses 14 physiological and 6 surgical variables for prediction of mortality and diseases. This model determines the mortality with the formula: the probability of death = -7.04 + (0.13 × the physiological value) + (0.16 × surgical value of severity) [24]. The estimation of physiological capability and surgical stress (E-PASS) consists of the preoperative risk score (PRS) (age, concurrent diseases, performance index and ASA) and the surgical stress score (SSS) (blood loss in relation to body mass, surgery time, skin incision size). It uses the universal risk score: 0.328 + (0.936 × PRS) + (0.976 × SSS) [25, 26]. Although these formulae are potentially able to predict the mortality in patients with hip joint fractures, it is quite difficult to use the complex formulae in real clinical practice. Moreover, the significant part of the data is not easy available. Such data are to be collected and analyzed.
In this study, the model for selection of surgical management of proximal femoral fractures considered six parameters: gender, age, a comorbidity category, potential risk of complications, ASA class severity, a fracture type.
Our simple nomogram for clinical estimation of the risk of complications at the background of concurrent diseases in patients with proximal femoral fractures with consideration of gender and age [16] was used in combination with screening of previous functional and physiological parameters for identification of patients with maximal risk of the complicated clinical course in the postsurgical period.
Our study identified five main clinical factors (age, gender, a comorbidity category, ASA severity class, and a fracture type as the predictors of postsurgical complications after proximal femoral fractures. These factors include the older age (65+), male sex, a comorbidity category 3+ concurrent diseases and ASA class 3-4 as more efficient indicators of development of postsurgical complications in comparison with other risk factors (the table 2).
The patients were successfully distributed into the subgroups of low, middle and high risk. The observed amount of postsurgical complications was most close to the data calculated with the nomogram.
Our study has some disadvantages – mostly because of the retrospective design. The cutoff values are based on the available models of risk prediction and exert the negative influence on the predictive efficiency of the final model. For example, our study did not include any patients with ASA classes 5-6. Future validation studies will help to investigate the possibilities of the alternative values, which can improve the model.
Our model uses the prime whole numbers for each parameter in patients with proximal femoral fractures. We think that the simple model has the highest applicability and practicability for daily clinical practice.
CONCLUSION
As result of the study, we developed a method for selection of surgical management for patients with proximal femoral fractures with use of the nomogram of the risk of postsurgical complications.
The combination of the parameters is used for selection of surgical management of proximal femoral fractures: gender, age, a comorbidity category (0 – no concurrent diseases, 1-2 concurrent diseases – average chronic condition, 3 and more concurrent diseases – multiple morbidity), the potential risk of complications, ASA severity class, a fracture type (intraarticular medial fractures or femoral neck fractures – subcapital, transcervical, basicervical; extraarticular lateral or trochanteric fractures – intertrochanteric, transtrochanteric) with use of the developed nomogram. If the combination of ASA class 4 with the potential risk of complications > 30 %, surgical interventions are not conducted; for ASA class 3 and the potential risk of complications < 30 % in medial and lateral fractures, low invasive osteosynthesis of the proximal femur is conducted; for ASA classes 1 or 2 with the potential risk of complications < 30 % in lateral and basicervical fractures, osteosynthesis is conducted; for medial fractures (subcapital and transcervical) – total hip replacement.
The nomogram for estimation of the risk of postsurgical complications in proximal femoral fractures allows fast predictor assessment without additional time-consuming calculations that is important for daily surgical clinical practice of traumatologist-orthopedist.
The offered method for selection of surgical management of proximal femoral fractures can identify the patients with higher risk of postsurgical complications and optimize the presurgical procedures for improvement in patient’s condition during surgery. Moreover, the method can be used as the tool for clarification of the potential and actual risk, allowing the objective comparison and testing the clinical outcomes.
Concerning the model for risk prediction, further studies are required for estimation of the external validation and comparison to identify the best predictor of postsurgical complications for each patient.
Information on financing and conflict of interests
The study was conducted without sponsorship.
The authors declare the absence of clear or potential conflicts of interests relating to publishing this article.