Risk factors associated with infection, amputation and mortality in patients submitted to infrainguinal arterial bypass. A retrospective review of 27 cases
(Portuguese
PDF version)
Luís Henrique Gil França,1 Henrique Jorge Stahlke Jr.,2 Mariana Tosta Garschagen,3 Carlos Frederico Rodrigues Parchen3
1.
M.Sc., Universidade Federal do Paraná (UFPR), Curitiba, PR, Brazil.
2. Associate Professor, Hospital de Clínicas, Universidade Federal do Paraná (UFPR), Curitiba, PR, Brazil.
3. Undergraduate Student of Medicine, Universidade Federal do Paraná (UFPR), Curitiba, PR, Brazil.
Correspondence:
Luís Henrique Gil França
Rua Coronel Dulcídio, 1189/1801
CEP 80250-100 - Curitiba, PR
Brazil
Phone: +55 (41) 343.0963
E-mail: luishgf@hotmail.com
ABSTRACT
Objective:
To evaluate the risk factors of infection after lower extremity revascularization and their association with amputation and mortality.
Method: The records of 241 surgeries in 220 patients who underwent surgical management of the ischemic lower extremity below the inguinal ligament during a period of 5 years were reviewed. We included elective or urgency procedures for treatment of chronic critical limb ischemia, limb threatening ischemia, acute limb ischemia and treatment of peripheral aneurysms of the lower limbs. The patients submitted to surgical treatment resultant from vascular trauma, previous infection in the affected limb, reoperation of the infected fields and death during the first seven postoperatative days were excluded.
Results: The incidence of infection was 12.27%, and the amputation
rate 29.62%. The risk factors which reached statistical significance
(P < 5%) were sex, smoking, previous procedures, reoperation,
indication of surgery and type of vascular graft (prosthesis). Male
sex, non-smoking patients and elective procedures had an odds ratio
of 15, 9 and 8, respectively. Patients with infection caused by
gram negative bacteria and those with vascular prosthesis were at
an increased risk for amputation of 26 and 16 (odds ratio), respectively.
No risk factor with statistic significance was associated with mortality.
The mortality rate was 14.81%.
Conclusion: The risk factors for infection, after lower extremity revascularization, are previous procedures, reoperation, indication of surgery (elective) and type of vascular graft (prosthesis). The risk factors that influence the amputation rate are type of bacteria (gram-negative strains) and use of vascular prosthesis. No risk factors of infection were associated with mortality in this study.
Key-words:
ischemia, surgery, lower extremity, infection.
Palavras-chave: isquemia, cirurgia, membros inferiores, infecção.
J
Vasc Br 2004;3(3):214-22
With the improvement of life-expectancy, as a result also of better prophylactic treatments and diseases control, we may expect an increase in the number of patients with chronic arterial occlusive disease of the lower limbs and, consequently, in the number of surgical revascularization of ischemic limbs.1
It is well established that the greater saphenous vein (GSV), either in situ, reversed, or transposed is the best arterial substitute in surgical revascularization of lower limbs, and it is the choice for infrainguinal arterial bypass precedures.2 The greater saphenous vein has the best adaptation in the human organism and is also the most resistant to infections.1 However, the GSV may be a varicose or sclerosed vein, preventing its use as an arterial bypass,2 and requiring a prosthesis to be used instead. It has been suggested that the increase in the number of artificial grafts may be associated with a higher incidence of infections.2
Vascular prosthetic graft infection is one of the most feared complications, leading not only to increased morbidity, but also to significant mortality.3,4 Effective strategies for the prevention and treatment of prosthetic infection pose a challenge to the surgeon. Morbidity and mortality associated with infection do not depend only on time factors, but also on the bacteria involved, graft site and treatment followed.5 The mortality rate for undiagnosed or not adequately handled infrainguinal arterial bypass infections are among 0 and 22%, and the amputation rate is among 8 and 53%.
The lack of data regarding the identification of risk factors in infectious
processes and their relation with morbidity and mortality in infrainguinal
arterial bypasses have motivated the authors to carry out this study.
Two-hundred and twenty patients who underwent infrainguinal revascularization
procedures carried out in the Vascular Surgery Discipline of the Hospital
de Clínicas of Universidade Federal do Paraná (UFPR) entered the
study.
The objectives of this study were: to assess the risk for infection in patients submitted to infrainguinal revascularization; to determine the association of risk factors of infection with the probability of amputation in patients who developed postoperative infection in infrainguinal arterial bypasses; and to determine the association of risk factors of infection with mortality in patients submitted to infrainguinal revascularization.
PATIENTS AND METHODS
Between February
1996 and February 2002, 220 records of patients submitted to a total of
241 surgical infra-inguinal revascularization procedures of lower limbs
were analyzed. Only infrainguinal arterial bypasses were studied. The
surgeries carried out are listed in Table 1.
 Table
1 - Types of infrainguinal arterial bypass grafts
 |
| Type |
|
n |
|
| Femoral-popliteal
|
above
knee |
89 |
| Femoral-popliteal
|
below
knee |
82 |
| Crossover
femoro-femoral |
|
14 |
| Femoral-fibular |
|
8 |
| Femoro-tibial
(anterior) |
|
10 |
| Femoro-tibial
(posterior) |
|
17 |
| Popliteal-distal
|
|
12 |
| Femoral-dorsalis
pedis |
|
1 |
| Femoral-trunk
tibio-fibular |
|
6 |
| Popliteal-profunda
femoral |
above
knee |
2 |
|
Grafts
used were ipsilateral GSV (186 - 77.18%), PTFE prosthesis (51 - 21.16%)
and Dacron (4 - 1.66%) prosthesis. Dacron, PTFE and reversed grafts
were implanted in the subfascial position.
Elective and urgent surgeries for the treatment of disabling claudication,
critical limb ischemia, acute limb ischemia and correction of peripheral
aneurysms of the lower limbs were included. The patients submitted to
surgical treatment of vascular trauma, previous infection in the affected
limb, reoperation in infected fields and death during the first 7 postoperative
days were excluded, as well as those whose records were not complete.
A protocol containing information about clinical records and the type
of infection was filled for each patient and approved by The Ethics
Committee of Research of the Hospital de Clínicas of UFPR.
Patients were assessed regarding type of treatment carried out and postoperative
evolution, both during the inpatient period and after inpatient discharge.
The follow-up period varied from 30 days to seven years. The assessment
followed the routine of admissions to the Hospital de Clínicas
of UFPR in what concerns anamnesis and physical examination. An assessment
through complementary tests was carried out in patients with suspect
or confirmed diagnosis of lower limbs ischemia. No patients presented
with renal disorder (creatinine level greater than 2.4 mg/dl or on dyalisis)
or clinical symptoms of malnutrition, although we have not carried out
an assessment of the level of plasma proteins, including albumin.
Epidural anesthesia was administered to the patients, and they received
1g of intravenous prophylactic cefazolin during anesthesia induction.
Criteria and results used to determine the presence of local infection
were approached according to the Recommended standards for reports dealing
with lower extremity ischemia: Revised version7 developed by the Ad
Hoc Committee of the Society for Vascular Surgery and of the International
Cardiovascular Society. Infections were graded according to the modified
Szilagyi criteria.8
Variables used in this study included: infection, age (under 70 years-old and over 70 years-old), sex, surgery (distal anastomosis above or below knee), diabetes mellitus, hypertension, smoking, chronic obstructive pulmonary disease, cardiopathy (arrhythmia, coronary failure and cardiac insufficiency), trophic lesion, previous procedures, type of vascular graft (PTFE and Dacron prosthesis or autologous vein), type of incision (continuous or interrupted), site of surgery (distal or inguinal), obesity, level of infection, time of infection (early < 30 days or late > 30 days), site of infection (inguinal or distal), amputation, death and type of bacteria (gram-negative or gram-positive).
The statistical analysis was performed in two parts: a descriptive one,9 in which basic descriptive statistics were determined, and an inferential one, based on logistic models,10 in which the statistical significance of risk factors related to the study variables was approached (described in Appendix) .
RESULTS
Twenty-seven out of 241 operations in 220 patients were diagnosed with postoperative infection, characterized in Table 2.
Table
2 - Features of the infected patients group
|
| Feature |
n |
|
| Age |
|
| Patients
under 70 years-old |
18 |
| Patients
over 70 years-old |
9 |
| Male |
25 |
| Smoking
|
21 |
| Hypertension
|
15 |
| Obesity |
3 |
| Diabetes
mellitus |
5 |
| Cardiopathy |
6 |
| Chronic
obstructive pulmonary disease |
3 |
| Trophic
lesion of lower limbs |
7 |
| Reoperation
and prosthesis extension |
10 |
| Previous
procedures |
11 |
|
As for previous procedures, seven were aortobifemoral, one external aorto-iliac, one external iliac, one iliac-iliac, one crossover femorofemoral, and two femoropopliteal above knee with prosthesis. Six out of seven patients with previous aortobifemoral surgery had been submitted to more than one surgical procedure.
With regards to indication of surgery, six patients were treating disabling claudication, 18 had critical limb ischemia and two underwent surgery for distal aneurysm correction. Two patients were submitted to urgent procedures, the remaining were elective interventions. Eleven patients were classified grade I infection under Szilagyi's classification, four patients were labeled grade II, six grade III, one grade IV, and five grade V. Grade I infections involved wound with erithema, lymphorrhagia and local cellulitis; Grade II presented fistula and abscess formation; Grade III had prosthesis exposure in six cases; there was one case of pseudoaneurysm in a Grade IV patient; and Grade V patients presented anastomosis with severe hemorrhage.
Types of surgeries and vascular grafts used in the infected group are listed in Table 3. Table
3 - Types of surgeries and grafts
|
| Surgery |
Graft |
n |
|
| Femoral-popliteal
above knee |
reverse
saphenous vein |
8 |
|
Dacron
prosthesis |
1 |
|
PTFE
prosthesis |
4 |
| Femoral-popliteal
below knee |
In
situ saphenous vein |
5 |
|
Reverse
saphenous vein |
2 |
|
PTFE
prosthesis |
1 |
| Crossover
femoro-femoral |
PTFE
prosthesis |
2 |
| Femoro-tibial
(anterior) |
In
situ saphenous vein |
1 |
| Femoro-tibial
(posterior) |
In
situ saphenous vein |
1 |
| Popliteal-distal
|
Reverse
saphenous vein |
2 |
|
PTFE grafts were implanted in above-knee femoropopliteal surgeries, and only one case received a below-knee prosthesis for the treatment of critical limb ischemia, because the patient could not undergo an autologous vein graft implant.
Continuous incisions were made in six surgeries with reverse greater saphenous vein. Interrupted incisions were accomplished in six surgeries with reverse greater saphenous vein and implant of PTFE graft. Infection occurred in sixteen patients with distal below-knee anastomosis, and in 11 patients with distal above-knee anastomosis.
Material for culture was collected from patients with suspect or confirmed
diagnosis of postoperative infection. They were treated according to
the antibiogram results. Most frequent bacteria were isolated and are
listed in Table 4. Specific cultures for Staphylococcus epidermidis
were not undertaken.
Table
4 - Bacteria identified
|
| Bacteria |
n |
|
| Staphylococcus
aureus |
8 |
| Staphylococcus
coagulase-negative |
6 |
| Streptococcus
pyogenes |
1 |
| Pseudomonas
aeruginosa |
6 |
| Proteus
sp. |
5 |
| Klebsiella
sp. |
1 |
| Escherichia
Coli |
1 |
| Acinetobacter
sp. |
1 |
| Enterobacter
cloacae |
1 |
| Serratia
sp. |
1 |
|
With regards
to the time variable, twenty-one patients (77.78%) presented with infections
within thirty days of the postoperative period, and six patients (22.22%)
after 30 days. In all patients labeled grade I and II of Szilagyi, infection
occurred within 30 days of the postoperative period, and cultures were
positive for Staphylococcus aureus, Streptococcus pyogenes,
Klebsiella sp., Escherichia coli, Pseudomonas aeruginosa
and Proteus sp. Patients infected with Pseudomonas aeruginosa
and Proteus sp. had sudden bleeding with anastomosis rupture.
Six patients were infected after 30 postoperative days and culture was
positive for: Staphylococcus aureus, Staphylococcus coagulase-negative,
Acinetobacter sp., Enterobacter cloacae, Serratia sp.
All of them underwent limb amputation.
Treatment followed in the infected group is listed in Table 5.
Table
5 - Treatment provided
|
| Treatment |
n |
|
| Antibioticotherapy
and bandage |
10 |
| Debridement,
bandages and antibiotic (prosthesis not removed) |
5 |
| Debridement
(prosthesis removed without revascularization) |
8 |
| Extra-anatomic
arterial bypass |
1 |
| Wound
coverage with skin graft |
1 |
| Ligature
of the graft |
1 |
| Correction
of pseudoaneurysm |
1 |
|
Evolution of the 11 patients labeled grade I under Szilagyi's classification was satisfactory. Two patients grade II of Szilagyi and one grade III also presented good evolution. A patient labeled grade III underwent successful surgical correction of pseudoaneurysm. Eight patients underwent amputation - six supracondylar procedures (four patients with grade III and two patients with grade IV); and two were infracondylar (patients grade III of Szilagyi). Four patients died as a result of shock (two cases), sepsis (one case) and cardiac insufficiency (one case). The amputation rate in the infected group was 29.62% (7.4% infracondylar and 22.22% supracondylar). This study observed that patients with grades I and II of Szilagyi had successful evolution with conservative treatment during the follow-up period. Patients labeled grades III, IV and V had a higher level of morbi-mortality. The mortality estimate of patients with postoperative infection was 14.81%.
Results
of statistical analysis
The rate of infection in infrainguinal arterial bypasses was 12.27%. The site of infection was the inguinal region in 20 patients (74.07%), and below-knee incisions in seven (25.93%) - the inguinal region was not dissected in two surgeries.
Variable: infection
Among possible risk factors, the following had the most statistical significance: reoperation, sex, smoking, indication of surgery, type of vascular graft, and previous procedures. Table 6 shows the effects of these factors on infections and the respective odds ratio estimates.
Table
6 - Odds ratio for risk factors
|
|
|
Confidence
interval of 95% |
| Variable |
Estimate |
Lower
limit |
Upper
limit |
|
| Reoperation
(Yes/No) |
4.884 |
1.720
|
13.865 |
| Sex
(Male/Female) |
14.853
|
2.570
|
85.849 |
| Smoking
(Yes/No) |
0.124
|
0.033
|
0.469 |
| Procedures
(Emergence/Elective) |
0.110
|
0.019
|
0.628 |
| Type
of graft (Prosthesis/Vein) |
3.626
|
1.283
|
10.251 |
| Previous
procedure (Yes/No) |
2.641 |
1.021 |
6.832 |
|
Sex was the most significant risk factor, with odds ratio of infection almost 15 times higher for male. The second higher risk factor was surgery indication, with odds ratio of infection almost nine times (=1/0, 110) higher for elective procedures. The third factor is associated with smoking patients, with odds ratio of infection almost eight times (=1/0, 124) higher for non-smoking patients (see Table 6 for other factors).
Variable: amputation
Only the 27 infected patients were considered in the analysis of the amputation variable, making possible to build only two models with only one significant risk factor. Among the risk factors analyzed, type of bacteria was statistically significant (see Table 7).
Table
7 - Effect of type of bacteria
|
|
|
Confidence
interval of 95% |
| Variable |
Estimate |
Lower
inferior |
Upper
limit |
|
| Bacteria
(Gram+/Gram-) |
0.038 |
0.004 |
0.407 |
|
Odds ratio of gram-negative bacteria was nearly 26 times higher. The probabilities of an infected patient undergoing amputation due to features of the risk factors selected in the logistic model are 0.6364% for gram-negative bacteria and 0.0625% for gram-positive bacteria.
Among possible risk factors, the type of vascular graft also showed to be statistically significant. The effect of this factor in amputation can be seen in Table 8.
Table
8 - Effect of type of graft
|
|
|
Confidence
interval of 95% |
Variable
|
Estimate |
Lower
limit |
Upper
limit |
|
| Type
of graft (Vein/Prosthesis) |
16.000
|
2.122
|
120.648 |
|
The type of vascular graft had an odds ratio of amputation nearly 16 times higher for the type of prosthesis. The probability of an infected patient undergoing amputation due to features associated to risk factors selected in the logistic model are 0.1111% for vein graft and 0.6667% for prosthesis.
Variable: death
Only the 27 infected patients were considered in the analysis of the death variable, probably the reason why no risk factor showed to be statistically significant. The probability of an infected patient dying is 0.1481 or 14.81%.
DISCUSSION
Infection is a dynamic process in which pathogenic microorganisms invade the body and tissues react to the germs and their toxins. The infection become a manifest disease only when the equilibrium between the host and the parasite is broken.11 The concepts of contamination and infection is very important, as the majority of surgical incisions is contaminated, but not infected, by the skin's natural flora.
It is reported that a load of 105 bacteria/g of tissue or biological fluid are necessary to a body to be considered infected.11 A classification of surgical wounds has been created in order to provide criteria for prevention of surgical infections. It revealed that there is a clear relation between the degree of contamination and the incidence of postoperative infection of surgical wounds.11 Procedures are considered to be clean when they are elective, primarily closed without drainage, in which respiratory, gastrointestinal, and genitourinary tracts are not entered and the aseptic factor has not been violated. In general, vascular surgical procedures are classified as clean, with an infection rate below 5%.11
The study of the risk factors for infection and their influence on the morbi-mortality rate of patients is important for the understanding of the physiopathology, prevention and treatment of postoperative complications, as surgical infections may cause physical incapacity and even death, besides extending the hospital stay and cost of hospitalization.11 Factors such as malnutrition, uraemia, use of chronic corticotherapy and chemotherapy, among others, are related to long-term wound healing and predisposition to infection. Sex, ulcer in the operated limb and reoperation are some risk factors associated with the increase in the number of vascular surgery infections.12
Several authors gathered records of their patients in order to try to find risk factors for infection in the postoperative period of infrainguinal arterial bypass surgeries. Reifsnyder et al. came to the conclusion that early reoperation (< 4 days) and presence of lymphatic fistula increased the risk for postoperative infection significantly.13 Nevertheless, other factors such as age, race, diabetes, timing of surgery and gangrene or ulcer did not demonstrate to have influence on the incidence of infections. Kent et al. report that postoperative infections in lower extremity revascularization surgeries increase the financial cost of patient care and that old age, postoperative venous stasis and obesity increased significantly the risk for infection.14 Treiman et al. consider the following factors as related to an increase in the incidence of infections: reoperations or reviews of previous arterial surgeries and the fact that all grafts were placed under the subcutaneous tissues.15 Wengrovitz et al. identified the following risk factors associated to a significant incidence of surgical wound complications in patients submitted to lower extremity revascularization with in situ GSV: chronic corticotherapy, ulcer in the ipsilateral limb and arterial bypass for dorsalis pedis artery; as associated factors he mentioned: female sex, use of continuous incision, diabetes mellitus and revascularization for limb salvage.16 The relation of these factors with vascular graft infection was not studied. Lorentzen et al. noticed that the infection developed only at the region of the inguinal incisions, and suggested that this procedure increases the risk for infection.17 Despite the high incidence of infection in the inguinal region, it was not considered a statistically significant variable in the present study, and it was not related with the risk for postoperative infection.
The following risk factors for infection in infrainguinal arterial bypasses
were identified (P < 0.05): sex, smoking, previous procedures,
use of prosthesis as arterial substitute, reoperations, and elective
procedures. Variables considered were: age (patients over 70 years-old)
and obesity, which have not demonstrated to be statistically significant
in the risk for infection. The null hypothesis was rejected in the univariate
analysis in some clinical variables, such as diabetes mellitus, hypertension
and cardiopathy; the same has not occurred in the multivariate analysis,
because in this case, patient's outcomes are analyzed together, while
in the first analysis variables are crossed in a one-to-one relationship,
having infection or not.
Smoking and male sex as risk factors in infranguinal arterial bypass surgeries may lead to bias in epidemiological studies, as the majority of patients submitted to revascularization surgeries are men and smoking. In the present study, the statistical analysis surprisingly showed that male and smoking patients were the most prone to develop infections; when it is generally accepted that the female sex is an independent risk predictor in arterial surgeries and that smoking impairs surgical wound healing, and these factors could be related to infections in the postoperative period.3,11,16 This may have been caused by the small number of patients in the infected group sample, characterizing an statistical error type I.
In a retrospective study, Chang et al. found only surgery time as a significant variable in the occurrence of infection, and concluded that two surgical teams working together in more time-consuming procedures could decrease the risk for postoperative infection. The author also mentions the importance of rigid antibiotic prophylaxis to lessen the risk for infection.12 Kent et al. and Wengrovitz et al. mention the surgical technique as an important variable to be considered.14,16 Kent et al. and Aguiar et al.14,18 stress that when a study about infection in infrainguinal arterial bypass is undertaken, it must be considered that patients are operated by different surgeons, with different teams and with the presence of residents, who are transient participants in the discipline, making difficult to analyze the surgical technique related with the incidence of infection.
Reoperations (other lower extremities revascularization surgeries for complementary care, reviews or extensions of previous vascular grafts) were also entered into this study, as it was known that manipulation of implanted prosthesis or new implants in the replacement of older facilitate the development of infections.19
Other authors, such as Reifsnyder et al. and Marsan et al. mentioned that a thrombosed prosthetic graft is a risk for infection of an adjacent patent graft, as well as those with lymphatic fistula and submitted to reoperation due to graft thrombosis or hematome in the surgical wound.13,20 In the present study, surgical procedures with distal anastomosis above and below knee were analyzed, and no difference was found concerning amputation, as well as statistical validity in the risk factor for infection. Calligaro et al., however, reports that the most distal the revascularization is, such as inframalleolar bypasses with history of infections, the higher the amputation rates are.21
Different properties are associated with different types of prosthesis
vascular grafts. Vein grafts have biological properties which inhibit
bacterial adherence, but they are prone to infections and rupture.1,22
Dacron grafts are not very resistant to infection and primary repair,
and the venous patch grafts present elevated rupture rates as a result
of mechanical failure of anastomosis. This study assessed surgeries
using both autologous veins and prosthesis, as well as the graft placement
in the subcutaneous tissue or subfascial position. A higher prevalence
of infection was demonstrated with the use of prosthesis (either Dacron
or PTFE) in infrainguinal arterial bypasses. This risk factor was considered
to be statistically significant (P < 0.05%).
The association of morbidity with infection depends on several factors.
Time of surgery and type of bacteria, specially gram-negative ones,
have great influence on the evolution of infected patients.23-25
In this study, the possibility of amputation was nearly 26 times higher
for patients infected gram-negative bacteria. Similar results have been
reported in the literature. Ouriel et al. report cases in which rupture
of arterial anastomosis and graft reinfection were more frequently associated
to infections caused by gram-negative bacteria and that the result of
infected patients treatment depends on type of bacteria.26
Geary et al. and Calligaro et al. say that the identification of Pseudomonas
sp. in affected sites deserves an aggressive treatment due to the
high virulence of this microorganism.23,27
With relation to time, the present study has not showed any association
of this variable with amputation and death rates, and no risk factor
was associated with mortality of patients infected after surgery. Similar
results were reported by Calligaro et al., who did not find any differences
in bacterial strains, but infections that occurred later on had higher
morbi-mortality.28
CONCLUSIONS
Risk factors for postoperative infections in infrainguinal arterial bypasses which showed to be significant were: reoperation, surgery indication (elective procedures), type of graft (PTFE and Dacron prosthesis), and previous surgery. Patients infected with gram-positive bacteria and using prosthesis as arterial substitute have higher probability of amputation. Risk factors significantly associated with the mortality of patients presented with postoperative infection after infrainguinal revascularization procedures were not observed.
REFERENCES
1.
Piano G. Infections in lower extremity vascular grafts. Surg Clin North Am 1995;75:799-809.
2. von Ristow A, Cury Filho JM. Aterosclerose obliterante periférica: tratamento cirúrgico das lesões abaixo do ligamento inguinal. In: Maffei FHA. Doenças vasculares periféricas. 2ª ed. Rio de Janeiro, RJ: Medsi; 1995. p. 489-519.
3. Coe DA, Towne JB. Infection control in lower extremity revascularization. In: Yao JST, Pearce WH. Progress in Vascular Surgery. Stamford, Connecticut: Appleton & Lange; 1997. p. 257-270.
4. Dellinger EP. Infecções cirúrgicas. Infecções cirúrgicas e escolha de antibióticos. In: Sabiston DC Jr. Tratado de cirurgia. As bases biológicas da prática cirúrgica moderna. 15ª ed. Rio de Janeiro, RJ: Guanabara Koogan S.A.; 1999. p. 248-263.
5. Fry DE, Marek JM, Langsfeld M. Infection in the ischemic lower extremity. Surg Clin North Am 1998;78:465-79.
6. Kikta MJ, Goodson SF, Bishara RA, et al. Mortality and limb loss with infected infrainguinal bypass. J Vasc Surg 1987;5:566-71.
7. Rutherford RB. Recommended standards for reports dealing with lower extremity ischemia: revised version. J Vasc Surg 1997;26:517-38.
8. Szilagyi DE, Smith RF, Elliot JP, Vrandecic MP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-33.
9. Bussab WO, Moretin PA. Medidas resumo. In: Bussab WO, Moretin PA. Estatística Básica. 5ª ed. São Paulo: Saraiva; 2002. p. 35-68.
10. Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. Applied Linear Regression Models. 3rd ed. U.S.A.: Irwin, Inc.; 1996.
11. Fontes B. Infecções em cirurgia. In: Aun F, Bevilacqua RG. Manual de cirurgia. São Paulo, SP: EPU; 1995. p. 193-212.
12. Chang JK, Calligaro KD, Ryan S, et al. Risk factors associated with infection of lower extremity revascularization: analysis of 365 procedures performed at a teaching hospital. Ann Vasc Surg 2003;17:91-6.
13. Reifsnyder T, Bandik D, Seabrook G, et al. Wound complications of in situ saphenous vein bypass technique. J Vasc Surg 1992;15:843-8.
14. Kent KC, Bartek S, Kuntz KM, et al. Prospective study of wound complications in continuous infrainguinal incisions after lower limb arterial reconstruction: incidence, risk factors, and cost. Surgery 1996;119:378-83.
15. Treiman GS, Copland S, Yellin AE, et al. Wound complications involving infrainguinal autogenous vein grafts: a current evaluation of factors determining successful graft preservation. J Vasc Surg 2001;33:948-54.
16. Wengrovitz M, Atnip RG, Gifford RRM, et al. Wound complications of autogenous subcutaneos infrainguinal arterial bypass surgery: predisposing factors and management. J Vasc Surg 1990;11:156-61.
17. Lorentzen JE, Nielsen OM, Arendrup H, et al. Vascular graft infection: analysis of sixty-two graft infections in 2,411 consecutively implanted synthetic vascular grafts. Surgery 1985;98:81-6.
18. Aguiar ET, Albers MTV, Langer B, Puech-Leão LE. Incidência de infecções comprometendo próteses arteriais. Rev Paul Med 1985;103:239-42.
19. Nevelsteen A, Suy R, Daenen W, et al. Aorto-femoral grafting factors influencing late results. Surgery 1980;88:642-53.
20. Marsan BU, Curl GR, Pillai L, et al. The thrombosed prosthetic graft is a risk for infection of an adjacent graft. Am J Surg 1996;172:175-7.
21. Calligaro KD, Veith FJ, Dougherty MJ, DeLaurentis DA. Management and outcome of infrapopliteal arterial graft infections with distal graft involvement. Am J Surg 1996;172:178-80.
22. Bandyk DF. Infection in prosthetic vascular grafts. In: Rutherford RB. Vascular Surgery. 5th ed. Philadelphia: W. B. Saunders Co.; 2000. p. 733-751.
23. Calligaro KD. Infrainguinal prosthetic graft infection. In: Cronenwett JL, Rutherford RB. Decision Making in Vascular Surgery. Philadelphia, PA: W. B. Saunders Co.; 2001. p. 208-211.
24. Cherry KJ, Roland CF, Pairolero PC, et al. Infected femorodistal bypass: is graft removal mandatory? J Vasc Surg 1992;15:295-303.
25. Taylor SM, Weatherford DA, Langan EM, Lokey JS. Outcomes in the management of vascular prosthetic graft infections confined to the groin: a reappraisal. Ann Vasc Surg 1996;10:117-22.
26. Ouriel K, Geary KJ, Green RM, DeWeese J. Fate of the exposed saphenous vein graft. Am J Surg 1990;160:148-50.
27. Geary KJ, Tomkiewicz ZM, Harrison HN, et al. Differential effects of gram-negative and gram-positive infection on autogenous and prosthetic aortic grafts. J Vasc Surg 1990;16:337-46.
28. Calligaro KD, Veith FJ, Schwartz ML, et al. Differences in early versus late extracavitary arterial graft infections. J Vasc Surg 1995;22:680-6.
Appendix
Description
of the statistical inference analysis
The
statistical inference was based on the construction of a logistic model10
for the study of risk factors associated with infection, amputation and
death in patients submitted to surgery. The regression logistic model
models P probability of a certain event, for example, death or
postoperative infection, as a function of risk factors based on the odds
ratio. The odds ratio is defined by
,that is, the ratio between the occurrence or not of a certain event.
For example, the ratio between death probabilities (P) and no-death
(1-P). A value for odds ratio higher than 1 means that the probability
of an event occurring is higher than not occurring. The odds ratio is
the reason between two ODDS, for example, the ratio between ODDS smoking
and non-smoking individuals, if the odds ratio value is higher than one,
this means that the smoking individuals ODDS is higher than ODDS of non-smoking.
The Stepwise procedure was followed in the model construction, by using
the Forward method, as it follows: variables were individually inserted
and their statistical significance was checked (P value). The variable
with the smallest significant P (less than 5%) was entered into
the model; after it was inserted, the same procedure was carried out for
the other variables, up to when there were not statistically significant
variables to be inserted in the model.
|
