Predictable factors of infection in patients with open fractures in lower limbs

Abstracts

134 pacientes com fraturas expostas dos membros inferiores dos tipos II, IIIA, IIIB e IIIC foram estudados prospectivamente entre fevereiro de 1998 e maio de 2000 no Instituto de Ortopedia e Traumatologia da Faculdade de Medicina da Universidade de São Paulo para determinar os fatores de risco de infecção previsíveis Todos os pacientes foram registrados em um protocolo no qual eles tiveram amostras de fragmentos ósseos iniciais coletadas para cultura bacteriana, avaliação de condições clínica e administração precoce de antibiótico. Durante a abordagem inicial, as causas das fraturas, o tempo de exposição da fratura e o local onde os primeiros socorros foram recebidos foram observados. Durante o debridamento inicial, o volume de transfusão sangüínea, a classificação clinica ASA, o tempo cirúrgico, o ferimento cirurgico e o tipo de estabilização esquelética foram observados. Uma análise univariável foi realizada para identificar os riscos pervisíveis estatisticamente significantes para o desenvolvimento de infecções, com os seguintes resultados: tempo de exposição da fratura (p=0.0201), local dos primeiros socorros (p=0.400), tipo de fratura (p=0.0130), classificação ASA (p=0.0005), volume de transfusão de sangue (p=0.0002) tipo de osso fraturado (p=0.0052), tipo de acidente (p=0.0450), ferimento cirúrgico (p=0.0024), estabilização esquelética (p=0.0446), cultura bacteriana positiva na admissão (p=0.5290) e cirurgias concomitantes (p=0.1867). As variáveis com associação significante com a infecção foram introduzidas em uma equação de regresão multivariada (modelo logístico) para identificar as com efeitos independentes dos outros fatores. O modelo logístico final foi obtido e demonstrou as probabilidades de infecção nas fratruras expostas estudadas. Os riscos relativos revelados no modelo logístico final foram : volume de transfusão sangüínea (mais do que 1 unidade) - 6.4;classificação ASA nível III - 5.2; fixação interna do osso (imediata) - 3.9; osso fraturado (femur) - 3.5 and ferimento aberto - 3.0

Fraturas expostas; infecção; Membros inferiores

ORIGINAL ARTICLE

Predictable factors of infection in patients with open fractures in lower limbs

Ana Lúcia Lei Munhoz Lima^I; Arnaldo Valdir Zumiotti^II; David Everson Uip^III; Jorge dos Santos Silva^IV

^IImmunologist - Assistant Doctor of the Orthopedics and Traumatology Department at the General Hospital of the São Paulo University - Medical School

^IIFull Professor of the Orthopedics and Traumatology Department at the General Hospital of the São Paulo University - Medical School, Head of the Traumatology, Hand and Microsurgery Discipline of Orthopedics and Traumatology Department at the General Hospital of the São Paulo University - Medical School

^IIIInfectologist - Full Professor of the Infectious and Parasitic Diseases of the Hospital das Clínicas of FMUSP. Director of the Infection Service of the Heart Institute of HC-FMUSP

^IVAssistant Doctor of the Orthopedics and Traumatology Department at the General Hospital of the São Paulo University - Medical School

^{Correspondence}

SUMMARY

134 patients with open fractures of lower limbs type II, IIIA, IIIB and IIIC were prospectively studied between February 1998 and May 2000 at the Orthopedics and Traumatology Institute ofthe São Paulo University Medical School, to determine predictable risk factors of infection. All patients were enrolled to a protocol where they had initial bone fragment samples collected for bacterial cultures, clinical conditions evaluation and early introduction of antibiotics. During the initial approach, causes of fractures, time of bone exposure and place of first medical care were observed.During the initial debridement, volume of blood transfusion, clinical ASA classification, surgical time, surgical wound and type of skeletal stabilization were observed. Univariate analysis was performed to identify predictable statistically significant risks to develop infection, with the following results : time of bone exposure (p=0.0201), place of first medical care (p=0.400), type of fracture (p=0.0130), ASA classification (p=0.0005), volume of blood transfusion (p=0.0002), fractured bone (p=0.0052), type of accident (p=0.0450), surgical wound (p=0.0024), skeketal stabilization (p=0.0446), positive bacterial culture in admission (p=0.5290) and concomitant surgeries (p=0.1867). The variables with significant association with infection were introduced into a multivariate regression equation (logistic model) to identify those with independent effect from other factors. The final logistic model was obtained and demonstrated the probabilities of infection in the open fractures studied. The relative risks revealed in the final logistic model were : volume of blood transfusion (more than 1 unit) - 6.4; ASA classification level III - 5.2; Internal fixation of the bone (immediate) - 3.9; fractured bone (femur) - 3.5 and open wound - 3.0.

Key words: Fractures; Open; Infection; Lower extremity

INTRODUCTION

The treatment of open fractures of the lower limbs has changed significantly in the last two decades. Although some basic concepts were described more than two centuries ago, the negligence in their correct application has led to unsatisfactory results.

Deasult introduced the management of the debridement of devitalized tissues in France in the XVIIIth century, and it was strictly used in Napoleon's campaigns to decrease the mortality level due to open fractures. Although the comprehension that the strict debridement, the fracture stabilization and the cutaneous investment repair are important factors for a successful treatment, the proper techniques of reconstruction are later events.

The open fracture treatment has gone through several stages, firstly aiming life preservation followed by leg preservation and by infection prevention. It is currently associated with the affected limb function preservation.

The search for predictable infection factors in this paper aimed to contribute to the most adequate management choice in each case, regarding the early cutaneous investment and the skeletal stabilization method.

CASES AND METHODS:

From February 1998 to May 2000, 245 patients were prospectively evaluated presenting an open fracture diagnosis in the lower limbs. They were attended at the Orthopedics and Traumatology Institute of the General of São Paulo University – Medical School. All the patients presenting a Type I open fracture, according to Gustilo and Anderson⁽¹⁴⁾ classification, and the ones that underwent an initial surgical toilet of the open fracture in other services were excluded from analysis. So, the final analysis evaluated 134 patients presenting Type II and Type III open fractures in the lower limbs. 114 patients (85%) were men, from 9 to 88 years old and the mean age was 32,9 years old.

The fractures were classified as Type II, Type IIIA, IIIB and IIIC ⁽¹⁴⁾. The cases implied 57 Type II fractures, 41 Type IIIA fractures, 30 Type III B and six Type III C fractures.

The most damaged bones were those of the leg and foot, in Type II fractures; the bones of the leg, in Type III A; the ones of the leg, femur and foot, in type IIIB and femur, tibia and ankle in type C fractures.

Most of the patients studied were victims of high –energy accidents such as: running over (32%), motorcycle accidents (29%), car accidents (17%), falls from height (11%) and wounds caused by guns.

All of the patients underwent a standardized evaluation protocol consisting of:

1 - Initial approach:

• As the patient arrived at the Emergency, it was obtained an osseous fragment to perform aerobe and anaerobe bacterial cultures through adequate means standardized by the Institution;

• A standardized antibiotic therapy associated with Clindamycin plus aminoglycoside was maintained for 14 days;

• ASA classification (American Society of Anesthesiologists)⁽¹⁾ was performed on the general impression of the patient during the preoperative period;

• The time of bone exposure, surgical time, volume of blood transfusion, type of skeletal stabilization after surgical debridement (plaster splint, skeletal traction, external fixation, internal fixation by plates or intramedullary nailing, cutaneous investment repair), conditions of the initial surgical wound (open or closed)

2 - Infection diagnosis

• The criteria to define osseous infections in patients' evolution followed the rules of the Center for Diseases Control and Prevention⁽¹⁰⁾, concerning superficial and deep incision infections, and postoperative osteomyelitis.

3 - Follow-up approach

• The patients underwent a daily orthopedic-infectious evaluation besides new debridements to collect osseous fragments for the culture whenever the infection diagnosis was done. The antibiotic therapy was settled according to the results and the observation time lasted until hospital discharge when the acute osteomyelitis treatment was finished.

RESULTS

In the evolution of open fractures treatment, infection signs which had been previously defined were observed in 40,30% of the patients (n=54). The distribution of the infection cases according to the type of fracture is shown in the table 1.

In order to verify the relationship between the presence or absence of infection at each of the possibly predictable infection risks, an univariate analysis was performed and the variable result was the presence of infection in the development of the open fractures treatment⁽⁶⁾.

The variables studied between the group with or without infection were: bone exposure time, place of first medical care, type of open fracture, positive bacterial culture in admission, ASA classification, volume of blood transfusion, concomitant surgeries, fractured bones, type of accident, surgical time, conditions of surgical wound after the first intervention, type of initial skeletal stabilization.

The Table 2 shows the results of the univariate analysis

To the variables of classification, the results are the descriptive levels of the probability test of Pearson's chi-squared. The descriptive level of probability for the bone exposure time refers to the medians' equality test.

By taking into account the statistical significance found in the univariate analysis and considering the really predictive infection variables, it was performed a multivariate analysis. After three adjustments in the regression equation (logistic model), the results are found in Table 3.

By the final regression equation (logistic model), it was calculated the relative risk (Odds-Ratio) to each variable.

The estimate relative risks for each variable are described in the Table 4.

DISCUSSION

The importance of several possibly predictive infection factors in the open fractures evolution has been studied by several authors for many years. Some predictive factors have always been considered as significant in papers, but others have been rejected during research due to the absence of proofs of their effectiveness^(8,9,13).

In this series, several probably predictive infection factors in open fractures were tested: exposure time, place of the first medical care after trauma, type of open fracture, positive bacterial culture in admission, ASA classification, volume of blood transfusion until the first surgery, type of trauma that caused the accident, concomitant surgeries, surgical time, open surgical wound after initial debridement and type of skeletal stabilization.

The statistical tests applied between the variables related to the result variable "infection" showed significance to: type of fracture, type of fractured bones, ASA classification, volume of blood transfusion, open surgical wound and skeletal stabilization.

The importance of the exposure time as a predictive factor of infection was demonstrated in several papers such as in Patzakis and Wilkins⁽²⁵⁾. They evaluated the predictive infection factors in 1104 open fractures and concluded that the ones of high risk are: lack of antibiotic therapy, exposure time longer than three hours, bacterial resistance to the antimicrobial used, a large tissular damage and tibia fractures. In fact, the authors considered the exposure time as the interval between the accident and the antibiotic therapy, and so they observed that if this interval is shorter than 12 hours but the patient is undergoing antibiotic therapy before the first three hours posttrauma, the predictive infection effect will be annihilated. In this classic paper, the author reports the importance of Type III fractures as an increasing risk to infection, but no numerical confirmation is presented.

The absence of exposure time significance in this sample is in accordance to Merritt's⁽²⁴⁾ findings. He also evaluated the risk factors of open fractures infection and this variable showed no importance. This author analyzed 70 patients presenting open fractures and a general infection level of 19%. The significant risk factors were the type of accident, the surgical time and confirming Patzakis and Wilkins⁽²⁵⁾ findings, the type II fractures with internal fixation.

Regarding the positive culture in admission as a predictive infection factor, all papers show its ineffectiveness, emphasizing that it shouldn't be performed.

However, some authors report the positive culture in admission as a significant factor in order to alert to the use of immediate internal fixation in these patients and the high risk of infection in such conditions⁽²⁴⁾.

The clinical severity scale of the patients, in this particular sample, according to the ASA classification, showed a high importance as an infection predictive and this fact has been frequently reported in various studies. In a study analyzing 53 open fractures of femur, the type of fracture, the severity of the patient, the delayed closing of the surgical wound and the immediate internal fixation⁽¹²⁾ were considered as predisposing to infection.

In another revision study, it was observed the reasons which lead the most severe cases to an increased infection risk. The hypotension generated by severe fractures with an extense injury of soft parts and the exposure time longer than six hours, leads to a considerable decrease of blood perfusion in muscles and bones besides a consequent oxygenation drop and insufficient intake of antimicrobials even when they are used in adequate doses. This intake decrease associated with the decreased phagocytosis common in acute traumas, affords highly favorable conditions to the bacterial growth⁽¹⁹⁾.

Maybe reflecting the severity of the patient and the fracture, the volume of blood transfusion is considered and attested by many authors as a predictive infection factor ^(26,27). At first, we found it surprising when this variable was considered in some papers as a more important factor than the bone exposure time, for example. But as this statement was constant in many papers, we decided to include it in our research about the predictive infection factors and we found a high statistical significance in the sample studied. Actually, by analyzing all the predictive factors found in our sample, and reevaluating the consequences of the severity of the trauma in the microcirculation and the bacterial risks of infection associated with the transfusion of red blood cells in patients that underwent a hip surgery and a general surgery, the importance of this variable as a predictive infection factor had to be reevaluated ^(3,4).

The type of accident that caused the fracture is frequently associated with the severity of the injury as well as to the severity of the patient. In all later studies in the literature, the high speed accidents are reported as the main causes of severe open fractures. Obviously, these later quotations show the changes in the types of accidents generating fractures according to the modern life style and technological advances that despite helping us may cause injuries like those. This fact frequently limits the comparison that could be made between earlier and later studies. We should take into account the progressive changes in trauma attack and consequently the severity o both fractures and patients^(24,28,29).

A study with 70 patients presenting open fractures showed a significant increase in infections when the injuries were generated by ground accidents, on farms, where the infection risks reach around 100% due to the nature and the quantity of inoculated microorganisms in the wound at the moment of the accident.

In the same sample, the importance of the surgical time as a predictive infection factor in postoperative fractures was analyzed taking into account that the patients with a surgical time longer than 121 minutes showed a higher infection risk. This datum was also found in our study. When we grouped patients in shorter intervals of surgical time, we could observe an increase of infection cases in those patients with longer surgical time⁽²⁴⁾

The biggest controversies in all consulted literature are related to the maintenance of the surgical open wound after the initial debridement and the type of skeletal stabilization chosen in that moment. Therefore, these two predictive infection factors in the evolution of the case will be discussed together.

In spite of the evolution of biomechanical concepts, surgical techniques, osseous fixation materials and antibiotic therapy occurred in the last 30 years, some authors haven't changed their minds about keeping the wound open after the initial debridement^(15,16). In his later studies, Gustilo reports the following paragraph: "If the surgeon is in doubt about the debridement to be performed in an open fracture, the wound shouldn't be closed and the safe rule is to keep it open." ⁽¹⁷⁾

By taking into account all the experience in treating patients that present severe open fractures for many years and using the confirmed data of this paper, we don't agree with this opinion, as we consider that if the wound is kept open for a long time, specially in extense injuries in soft parts, this can be one of the highest risk factors to acute postoperative osteomyelitis caused by intra-hospitalar microorganisms and multiresistant antimicrobials that appear in these cases⁽²³⁾. We know that acute posttrauma osteomyelitis is the main cause of non-primary amputations in open fractures and when it is chronicle it shows a high morbidity rate leading the patient to permanently disabling injuries or to extremely long treatments with discouraging results⁽²⁰⁾. We consider that the cutaneous investment repair should be done early, when debridement has controlled the necrotic or infected tissues.

This opinion meets several other authors' who found the same importance in keeping the wound open for the development of infections in the evolution of exposed fractures.

Some studies report that tibia fractures treated with open wounds present a high level of complications compared to those treated with an early investment repair. In patients treated with the open wound the mean time for the consolidation of the fracture was nine months, with a mean hospitalization time of 45 days, osteomyelitis incidence of 41% and 29% of amputations. For those treated with an early investment, the mean time for the consolidation of the fracture was five and a half months, with a mean hospitalization time of 23 days, osteomyelitis incidence of 5% and 5% of amputations⁽²⁾.

Godina reinforced the findings reported in the publication above, demonstrating the percentage of infections and other complications in comparison to other three groups of patients with fractures of the borders and musculocutaneous investment performed in different time. The patients that underwent an early investment of the fractures of the borders (before 72 hours) presented 1,5% of infection, those with a delayed investment of fractures (between 72 hours and three months) presented 17,5% of infection⁽¹¹⁾.

Other authors reported that maybe the adequate management and the early closing of soft parts injuries are more important to prevent infections than the osseous fixation techniques applied⁽¹²⁾.

In a study performed in our Institution, the authors showed the use of microsurgical flaps in the treatment of open tibia fractures type III and, when this technique was applied in Emergency, there wasn't any infection development. In patients presenting a delay in the musculocutaneous repair, the osteomyelitis incidence was 18%⁽²⁹⁾.

As it was observed, most of these studies were performed with open tibia fractures and type III fractures. Now we know, and this is part of our findings, that the leg (tibia and femur), tibia and femur fractures type III represent a great set of predictive infection factors in treatment evolution. These fractures need particular care from the moment the patient arrives at the hospital, and all the actions should be directed in order to prevent from complications and infections.

Especially in these fractures, the initial orthopedic procedure as well as the best opportunity to perform the musculocutaneous investment have been questioned. The controversy is in applying the immediate internal fixation by external fixation or using external fixation during a period and after that converting it into internal fixation.

Merritt⁽²⁴⁾ considered some infection risk factors in open fractures settling the match of positive culture in admission, type III fracture and immediate internal fixation as the worst prognosis for complications.

Other authors have assumed very interesting positions, which we have shared with, as a conscious indication of immediate internal fixation, especially by intramedullary scapus, taking into account the known predictive infection factors and the surgeon's experience. There isn't a pre-settled rule and the case will determine the rule^(5,19).

In spite of these considerations, some authors have shown good results by using immediate internal fixation through intramedullary scapus. In a review study of 495 tibia fractures treated with this technique, very acceptable infection levels were observed: 3,8% for type II, 9,5% for type III. These authors also support that the opportunity and the quality of the cutaneous investment are more important to prevent from infections than the fixation technique⁽⁷⁾.

Regarding the fixation techniques, the most evident infection risk in the fractures evolution mentioned by several authors is the internal fixation with intramedullary scapus preceded by external fixation during a certain time.

Some reports show 44% of infection with this conversion procedure from external to internal fixation, after open fractures⁽²⁸⁾.

In this paper, it was confirmed that the patients that underwent early internal fixation with intramedullary nailing presented higher infection levels than those who used external fixation did. However, we couldn't associate the cutaneous investment in these cases.

After an individual analysis of the predictive infection factors, we performed a multivariate analysis by logistic regression aiming to determine the independent variables in predicting infection in open fractures evolution and to settle a logistic model to antecipate the patients with the highest infection risks.

In the literature, there is only one logistic model for infections in open fractures and it doesn't help us due to the fact that it was developed based on two Israel wars, in 1973 and 1982. Some analyzed variables are too particular for the war wounds and cares. As it is the only model, it is worth showing the independent predictive infection variables usually found in the war of 1973: multiple surgeries, amputations, injuries in several body sites and open drain systems. In 1982, the independent infection variables particularly for fractures were: multiple surgeries, drains, loss of soft parts and blood transfusion⁽²⁷⁾.

In the multivariate analysis performed in the present series, we found a high relative risk in the variables: ASA classification, transfusion of blood volume, type of fractured bone, wound kept open and type of skeletal stabilization.

CONCLUSION

Through these results, we could gather the situations that provided the highest infection levels in open fractures evolution in the sample studied. So we could transcribe that the patients presenting the most serious clinical severity, the most severe fracture, with leg and femur bones fractured and specially with type III fractures, which need a transfusion in the first surgery, should be carefully evaluated regarding the cutaneous investment and the skeletal stabilization method.

REFERÊNCIAS BIBLIOGRÁFICAS

Work performed at Orthopedy and Traumatology Department - FMUSP

• 1. American Society of Anesthesiologists. New classification of physical status. Anesthesiology 24:111-115, 1963.
• 2. Byrd HS, Spicer TE, Cierney G. Management of open tibial fractures. Plastic Reconstr Surg 76:719-730, 1985.
• 3. Carson JL, Altman DG, Noveck H, Sonnenberg FA, Hudson JI, Provenzano G. Risk of bacterial infection associated with allogeneic blood transfusion among patients undegoing hip fracture repair. Transfusion 39:694-700, 1999.
• 4. Chang H, Hall GA, Geerts WH, Greenwood C, McLeod RS, Sher GD. Allogeneic red blood cell transfusion is as independent risk factor for the development of postoperative bacterial infection. Vox Sang 78:13-18, 2000.
• 5. Chapman MW, Olson SA. Open fractures. In: Rockwood CA, Green DP. Fractures in adults. 4^thed. Philadelphia, Lippincott-Raven, 1996. p.305-352.
• 6. Collet D. Modelling binary data. London, Chapman and Hall, 1991.
• 7. Court Brown CM, Keating JF, McQueen MM. Infection after intramedullary nailing of the tibia. Incidence and protocol for management. J Bone Joint Surg Br 74:770-744, 1992.
• 8. Dellinger EP, Miller SD, Wertz MJ. Risk of infection after open fracture of the arm or leg. Arch Surg 123:403-404, 1990.
• 9. Fazzi A, Camanho GL, Tamari J, Tamaoki MTT, Otani RI. Fraturas expostas, análise de 364 casos. Rev Bras Ortop 13:186-188, 1978.
• 10. Garner JS. CDC guideline for prevention of surgical wound infections. Infect Control 7:193-200, 1985.
• 11. Godina M. Early microsurgical reconstruction of complex trauma of the extremities. Plastic Reconstr Surg 78:285-292, 1986.
• 12. Green A, Trafton PG. Early complications in the management of open femur fractures: a retrospective study. J Orthop Trauma 5:51-61, 1991.
• 13. Gustilo R, Simpson L, Nixon R, Ruiz A, Indeck W. Analysis of 511 open fractures. Clin. Orthop 66:148-154, 1969.
• 14. Gustilo R, Anderson J. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones. J Bone Joint Surg Am 58:453-458, 1976.
• 15. Gustilo R, Mendoza R, Williams D. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. J Trauma 24:742-746, 1984.
• 16. Gustilo R, Gruninger R, Davis T. Classification of type III (severe) open fractures relative to treatment and results. Orthopedics 10:1781-1788, 1987.
• 17. Gustilo R. Management of open fractures. In: Gustilo R, Gruninger R, Tsukayama D. Orthopaedic infection, diagnosis and treatment. Philadelphia, Saunders, 1989. p.87-117.
• 18. Hughes SPF. Antibiotic penetration into bone in relation to the immediate management of open fractures: a review. Acta Orthop Belg 58 (Suppl 1):217-221, 1992.
• 19. Kaltenecker G, Wruhs O, Quaicoe S. Lower infection rate after interlocking nailing in open fractures of femur and tibia. J Trauma 30:474-479, 1990.
• 20. Koval KJ, Meadows SE, Rosen H, Silver L, Zuckerman JD. Posttraumatic tibial osteomyelitis: a comparison of three treatment approaches. Orthopedics 15:455-460, 1992.
• 21. Kreder HJ, Armstrong P. The significance of perioperative cultures in open pediatric lower-extremity fractures. Clin Orthop 302:206-212, 1994.
• 22. Lee J. Efficacy of cultures in the management of open fractures. Clin Orthop 339:71-75, 1997.
• 23. Lima ALLM, Zumiotti AV. Aspectos atuais do diagnóstico e tratamento das osteomielites. Acta Ortop Bras 7:135-141, 1999.
• 24. Merritt K. Factors increasing the risk of infection in patients with open fractures. J Trauma 28:823-827, 1988.
• 25. Patzakis MJ, Wilkins J. Factors influencing infection rate in open fracture wounds. Clin Orthop 243:36-40, 1989.
• 26. Polk H. Factors influencing the risk of infection after trauma. Am J Surg 165(suppl):2-7, 1993.
• 27. Simchen E, Raz R, Stein H, Danon Y. Risk factors for infection in fracture war wounds (1973 and 1982 wars, Israel) . Mil Med 156:520-527, 1991.
• 28. Yokoyama K, Itoman M, Shindo M, Kai H. Contributing factors influencing type III open tibial fractures . J Trauma 38:788-793, 1995.
• 29. Zumiotti AV, Ohno PE, Guarnieri M. Tratamento das fraturas expostas da tíbia grau III com emprego de retalhos microcirúrgicos. Acta Ortop Bras 2:13-18, 1994.

Publication Dates

• Publication in this collection
  24 June 2004
• Date of issue
  Mar 2004

History

• Accepted
  16 Feb 2004
• Received
  31 July 2003