PROGNOSIS OF INFECTIOUS COMPLICATIONS IN SURGICAL SITE AFTER OSTEOSYNTHESIS OF LONG BONES Tkachenko A.N., Ekhsan-Ul-Khak, Korneenkov A.A., Kushnirchuk I.I., Rankov M.M., Khromov A.A., Boychenko A.V.
North-Western State Medical University named after I.I. Mechnikov,
Kirov Military Medical Academy,
Saint Petersburg State University, Saint Petersburg, Russia
PROGNOSIS OF INFECTIOUS COMPLICATIONS IN SURGICAL SITE AFTER OSTEOSYNTHESIS OF LONG BONES
Fractures of the long bones of the extremities take the leading place in the injury rates in the last decades. According to the different authors’ data, their proportion is from 16.7 to 49.8 % among all locomotor system injuries [1, 2]. Moreover, opened fractures are verified in 10-18 % of cases of all fractures of the long bones (LB) [3, 4].
Currently, the rate of early postsurgical local infectious complications after metal osteosynthesis (MOS) is 2-12 %, reaching 55.9 % for opened fractures of the leg bones [5-8]. Deep surgical site infection (SSI) is identified in 1.3-4 % of cases, with its rate up to 22.6 % in patients with complex opened fractures of the tibia [8, 9, 10].
During the recent years, the great amount of publications relating to the issues of prediction of local purulent complications in traumatology and orthopedics has appeared [11, 12, 13]. Also the literature includes some findings relating to predicting the infectious complications after surgical treatment of fractures of the long bones [14, 15]. However there is not any uniform opinion about types and significance of prediction criteria, and the findings of possibilities of such prediction are contradictory [16, 17]. As result, investigation of the issues of development of techniques for predicting the local infectious complications in LB osteosynthesis can be the actual problem of scientific medical researches.
Objective – based on development and use of prognosis and prevention of surgical site infection, to improve the outcomes of metal osteosynthesis in diaphyseal fractures of long bones.
MATERIALS AND METHODS
The study was conducted on the basis of the approval from the bioethical committee of North-Western State Medical University named after I.I. Mechnikov and corresponded to WMA Declaration of Helsinki - Ethical Principles for Medical Research Involving Human Subjects 2000 and the Rules for Clinical Practice in the Russian Federation confirmed by the Order of the Health Ministry of Russia, June 19, 2003, No.266. 347 patients received metal osteosynthesis of the long bones of the extremities in the clinical facilities of the department of traumatology, orthopedics and military field surgery of North-Western State Medical University named after I.I. Mechnikov (hereafter, the clinic) in 2011-2016. The patients were distributed into two groups: retrospective (230 persons who received MOS in 2011-2014) and prospective (117 persons who received MOS in 2015-2016).
The retrospective group the patients with mean age of 56.8 ± 18.2 (18-90).
The postsurgical mortality was 2.6 % (6 cases). 224 were discharged from the clinic. The postsurgical results could not be evaluated in 45 (20.1 %) patients 12 months after surgery because of lost connection or death. These patients were excluded from the study. During one year, the long term results were estimated in 179 (79.9 %) patients (the table 1).
Table 1. Characteristics of patients whose data were analyzed in the study
|
Comparison parameters |
Retrospectivegroup |
Prospectivegroup |
|
Mean age, years |
56.8 ± 18.2 |
54.2 ± 14.4 |
|
Gender: | ||
|
men, absolute (%) |
74 (41.3) |
48 (41.0) |
|
women, absolute (%) |
105 (58.7) |
69 (59.0) |
|
Fracture location: | ||
|
humerus, absolute (%) |
57 (39.6) |
43 (36.7) |
|
forearm, absolute (%) |
28 (19.4) |
16 (13.7) |
|
hip, absolute (%) |
24 (16.7) |
20 (17.1) |
|
leg, absolute (%) |
35 (24.3) |
38 (32.5) |
|
MOS type: | ||
|
external, absolute (%) |
90 (50.3) |
55 (47.1) |
|
intramedullary, absolute (%) |
52 (29.1) |
39 (33.3) |
|
extrafocal, absolute (%) |
32 (17.8) |
21 (17.9) |
|
others, absolute (%) |
5 (2.8) |
2 (1.7) |
|
Surgical site infection within 12 months, absolute (%) |
35 (19.6) |
11 (9.4) |
|
superficial, absolute (%) |
25 (14.0) |
8 (6.8) |
|
deep, absolute (%) |
10 (5.6) |
3 (2.6) |
Note: MOS – metal osteosynthesis; SSI – surgical site infection.
The stages of the study:
The stage I. The retrospective study. The analysis included the data of 179 patients with osteosynthesis of the long bones of the extremities after diaphyseal fractures. The information on 144 patients with metal osteosynthesis without infectious complications (the group 1) was compared to the patients with local infectious complications within one year after the surgery (the group 2, 35 cases).
1. The comparative analysis of clinical and anamnestic data.
2. Identification of significant predictive factors.
3. Estimation of information capacity of the factors.
4. Estimation of diagnostic efficiency of the model.
5. Development of preventive measures.
The Stage II. The prospective study. The program for SSI prediction was tested in 117 patients after metal osteosynthesis for fractures of the long bones.
1. Estimation of SSI prognosis.
2. Identification of the risk group of local infectious complications.
For the sum of +14 points and more, the risk of SSI was estimated as low, the prognosis – as favorable. The individual preventive measures were not conducted.
For the sum of -14 and lower, the risk was high, and the prognosis – unfavorable. The individual preventive procedures were conducted.
For the sum within the range from -14 to +14, the risk was moderate, the prognosis – unclear. The individual preventive procedures were conducted.
3. Estimation of the outcome for the minimal period of 12 months after surgery was performed.
The comparative analysis of two subgroups in the retrospective cohort was conducted. The risk factors of local infectious complications after LB metal osteosynthesis were identified. The significance of each factor was estimated in quantitative equivalent for creation of the mathematical prediction model of SSI. The efficiency of the developed technique for prediction and prevention of local infectious complications after MOS was estimated in 117 patients in the main (prospective) group who were operated in 2015-2016. The patients in the retrospective and prospective groups were comparable according to gender signs, the age and fracture location (the table 1).
The presurgical preparation and prevention of SSI was conducted in the retrospective group with use of the standard technique. For these 179 patients, osteosynthesis was conducted in 90 (50.3 %) cases, intramedullary one – in 52 (29.1 %) patients. Extrafocal osteosynthesis was conducted in 32 (17.8 %) cases. Other types of MOS (pins, wire, screws and others) were used in 5 (2.8 %) cases. Within one year after the surgery, some local infectious complications were confirmed in 35 (19.6 %) cases, with deep infection in 10 (5.6 %) patients. Both the superficial infection of the incision and the deep infection in the surgical site were considered.
The retrospective group was divided into two subgroups: I – clinical observations during 12 months after the surgery, without SSI (144 patients), II – the patients with superficial and deep infection in the surgery zone (35 patients).
The analysis included 85 parameters of the patient’s condition and the data of objective, laboratory and instrumental examination. They included the information on the general and local statuses of the patient (gender, age, concurrent pathology, body mass index and others), surgical interventions (surgery time, intrasurgical blood loss, type of intervention and others). Some laboratory and instrumental examinations were investigated separately.
The SSI prediction model used the factors with statistically significant (p < 0.05) association with the outcome, and the factors, which association with the outcome was confirmed by other studies [18-21]. P value was estimated with Openepi.com according to J.L. Fleiss et al (2003) [22].
The comparison of the data in two subgroups of the retrospective study was realized with the method of sequential analysis by A. Wald (1945) with modification by E.V. Gubler and A.A. Genkin (1973) [23, 24]. Both the risk factors of SSI and their retrospective quantitative equivalent were identified. On the basis of these retrospective rating values of SSI risk, the prediction model of the postsurgical course after fractures of the long bones was developed.
For making a solution, the summed index of prognosis was compared to the borderline values which were calculated with the formula:
The threshold A = 10 × ln ((1 - α)/β),
The threshold B = 10 × ln (α/(1 - β)),
where α and β – the errors of the first and second types [24]. The error of the first type α was a false prognosis of favorable time course of the postsurgical period, without development of local purulent complications, when the patient had SSI. The error of the second type β was incorrect identification of unfavorable confirmation of unfavorable course of the postsurgical period, without development of local purulent complications, when the patient had no SSI. The values of α and β were 0.2 (20 %).
The patients of the prospective group received the presurgical preparation with consideration of SSI prognosis.
RESULTS
Among the variety of the clinical and anamnestic parameters of different stages of the treatment (presurgical, intrasurgical, postsurgical), 18 ones were selected, which had some differences in the groups with favorable postsurgical course and with SSI. These parameters were the predictive criteria.
The predictive value in relation to infection risk in the surgical site includes the gender, the age, concurrent diseases, time of year, type and location of a fracture, duration of postsurgical period, risk of anesthesia (presurgical criteria). The intrasurgical predictive factors include the findings of surgery time, intrasurgical blood loss, a type of metal osteosynthesis and order of priority in the surgery room. The criterion of postsurgical prognosis is regimen of the patient.
As the example of calculation of the coefficient of prognosis, we give the data on distribution of the patients with consideration of fracture location (the item 9, the table 3) as one of the factors of SSI risk (the table 2).
Table 2. Distribution of patients with fractures of long bones with consideration of fracture location
|
Fracture location |
Number of cases in postsurgical period |
p-level | |||
|
without complications |
with SSI | ||||
|
àáñ. / abs. |
% |
àáñ. / abs. |
% | ||
|
Humerus |
57 |
39.6 |
6 |
17.1 |
0.006 |
|
Forearm |
28 |
19.4 |
7 |
20.0 |
0.47 |
|
Hip |
24 |
16.7 |
8 |
22.9 |
0.196 |
|
Leg |
35 |
24.3 |
14 |
40.0 |
0.031 |
|
TOTAL |
144 |
100.0 |
35 |
100.0 |
|
The table 2 shows 39.6 % of the patients with uncomplicated postsurgical course after humerus fractures, and 17.1 % (2.3 times lower) of the patients with humerus fracture as result of SSI. The reverse situation was observed in fractures of the leg bones. Among the patients with favorable postsurgical period, such cases were 24.3 %, among the patients with infectious complications – 40 %.
The statistical analysis showed that the proportions of the patients with and without complications had some significant differences in the patients with various location of a fracture (chi-square test = 14.206, degrees of freedom = 3, p = 0.0024). Therefore, the factor of fracture location was considered in development of the mathematic model of SSI prognosis.
So, 18 predictive criteria were selected from the variety of the parameters (the table 3). 12 criteria were identified in the presurgical period, 4 – during the intervention, 1 – in early postsurgical period.
Table 3. Structure of weight coefficients of criteria for presurgical prediction of SSI in patients with fractures of long bones
|
|
Number of cases (%) |
p-level |
Correlation index |
Prediction coefficient | |
|
Without complications |
SSI | ||||
|
1 |
2 |
3 |
4 |
5 |
6 |
|
Before surgery | |||||
|
1. Gender: |
|
|
0.044
|
|
|
|
2. Age, years: |
|
|
|
|
|
|
Concurrent pathology: |
|
|
|
|
|
|
8. Fracture type |
|
|
|
|
|
|
9. Fracture location |
|
|
|
|
|
|
10. Fracture type |
|
|
|
|
|
|
11. Time of surgery: |
|
|
|
|
|
|
12. Presurgical bed-day |
|
|
|
|
|
|
13. Anesthesia risk, ASA |
|
|
0.0009 |
|
|
|
Surgery | |||||
|
14. Metal osteosynthesis type |
|
|
|
|
|
|
15. Sequence in surgery room |
|
|
|
|
|
|
16. Surgery time |
|
|
|
|
|
|
17. Intrasurgical blood loss |
|
|
|
|
|
|
Postsurgical period | |||||
|
18. Postsurgical mode |
|
|
0.0002 |
|
|
After formation of the complete list of the predictive values, the index of the ratio and the coefficient of prognosis were calculated.
The correlation index was the quotient between the incidence of a sign in the group of the patients with favorable postsurgical course and the incidence in the patients with SSI. The prognosis coefficient was the natural logarithm (ln) of the correlation index, which was increased 10 times for comfortable calculations. As result, the prognosis coefficient was +8.5 for the humerus fracture, -0.5 for forearm fractures, -3.0 for femoral bone fractures, -5.1 for leg fractures. It allowed making the conclusion about a degree of increasing risk of SSI in case of a fracture location in the lower extremity.
Subsequently, all prognosis coefficients (known at the moment of the examination) were summed. The result was the summed prognosis index (PI). This parameter was estimated at various stages of the examination and treatment of the patient. Before surgery – with 13 items, with consideration of intrasurgical data – with 17 items, in early postsurgical period – with consideration of all 18 items.
Therefore, if the total index of PI was at the level of the cutoff value +14 and more, then, with more than 80 % possibility, the favorable postsurgical course without local purulent complications could be predicted. If PI total index was less than -14, then SSI could be anticipated. The table 3 shows the full list of the prediction criteria of local purulent complications.
The most significant criteria with the maximal range between positive and negative values of the prognosis coefficient were location of fracture, its type (opened or closed), surgery time, intrasurgical blood loss.
The estimation of the diagnostic efficiency of the model (according to the data of the retrospective group) showed that sensitivity (Se) was 94.3 % (80.8 % - 99.3 %), specificity (Sp) – 97.9 % (94.03 % - 99.57 %).
The positive predictive value of the test was 91.7 % (78.16 % - 97.13 %), the negative predictive value – 98.6 % (94.83 % - 99.63 %).
The offered technique for predicting SSI identifies the risk group in patients. The results were estimated in the group of prospective study including 117 clinical cases. Such risk was virtually supposed in 23 (19.7 %) of 117 patients. All 23 patients (the high risk group of SSI) and 31 (26.5 %) patients with unclear risk received the complex of the special preventive measures (local, general and antibiotic prevention). The general prevention measures included the prediction of SSI at the stage of presurgical examination; clarification of development of SSI with consideration of intrasurgical data; preparation of cardiovascular system: water-electrolytic balance correction, correction of cardiac rhythm disorders; correction of carbohydrate metabolism disorders; correction of intracellular hemostasis; respiratory preparation; traffic optimization in the surgery room. The local prevention measures included the following procedures: ultrasonic examination of the surgical site; puncture examination in a case of hematoma formation (according to the ultrasonic examination); the control of discharge through the drains; bacteriological examination of discharge and puncture material; magnetic and laser therapy for the surgical site.
After these measures, SSI was confirmed in 11 (9.4 %) patients after MOS for long bones fractures among 23 cases with high risk of purulent and inflammatory complications. Deep SSI developed in 3 (2.6 %) patients.
DISCUSSION
The recent literature includes the well described techniques for MOS-associated infection prediction. However there is not any uniform technique. Analyzing the prediction criteria of local infectious complications after surgery for trauma including LB fractures, many authors consider the time from injury to surgery as the main factor [20, 21, 25]. Some authors believe that the SSI rate depends on the osteosynthesis type. So, I.Yu. Ippolitov (2016) report on the higher rate of infections complications (11.5 %) after LB metal osteosynthesis as compared to intramedullary constructs (3.1 %) [7].
A.M. Miromanov et al. (2017) reported on their study of 163 fractures of the long bones with infectious complications. They offer to make the prognosis of infection in the surgical site (including chronic posttraumatic osteomyelitis) after MOS both in early and late periods of traumatic disease on the basis of genetic predisposition. Identification of the genotype -589Ò/Ò of IL-4 gene and the genotype -308À/À of TNFα gene is considered by these authors as the informative value of SSI development after fractures of the long bones [15].
Some authors consider some factors promoting the development of local complications in treatment of LB fractures: age of patients (> 60), adiposity, concurrent diseases (diabetes mellitus, decompensated pathology of cardiovascular and respiratory systems, presence of foci of silent infection), alimentary protein insufficiency, fracture location in the distal parts of the extremity, immune system dysfunction and others [14, 16, 17, 25]. But these researchers do not report on the significance of the above-mentioned prediction criteria. Mostly, the mentioned factors characterize either general condition of the patient or organizational parameters reflecting the availability of specialized orthopedic and traumatological care.
These studies have a number of the disadvantages. The authors do not give any recommendations for practical use of information on presence of one or other criterion in the patient. In some studies, the specialists try to estimate the risk of SSI with use of their scoring system, which considers the quantitative equivalent of each risk factor and, correspondingly, the total value, which allows separating the risk group in relation to local infectious complications. However the information on testing the offered systems and algorithms in prospective studies is scarce.
CONCLUSION
Therefore, the practical use of mathematical prediction of SSI and administration of the preventive measures for patients of the risk group has resulted in the decrease in the SSI rate after metal osteosynthesis of the long bones from 19.6 to 9.4 % (two times). The rate of deep infection also has decreased two times: from 5.6 to 2.6 %. The clinical approbation of the program for prediction and prevention in the patient of the prospective group confirms the correctness of choice of the risk factors of infection in the surgical site after MOS of long bones.
The rate of fractures of the long bones does not demonstrate any stable trend to decrease. Also the risk of postsurgical infectious complications after treatment of fractures of the long bones remains. As result, it is obvious that separation of the group of patients with high risk of SSI, and administration of preventive measures prevent the development of purulent complications in the surgical site. The practical administration of the offered technique for prediction of SSI after metal osteosynthesis of the long bones concludes in summing the prediction coefficients and does not require special training or equipment. The results show that the use of the modern organizational approaches for patients requiring for surgical treatment of fractures of the long bones allows preventing the development of local infectious complications in some cases.
Information about financing and conflict of interests
The study was conducted without sponsorship.
The authors declare the absence of any clear and potential conflicts of interests relating to publication of this article.