The consequences of an outbreak of multidrug-resistant Pseudomonas aeruginosa among patients treated for critical leg ischemia
Article Outline
Objective
This retrospective matched case-control study evaluated the consequences of multidrug-resistant Pseudomonas aeruginosa (MDR Pa) in critical leg ischemia (CLI) patients treated with infrainguinal bypass surgery (IBS).
Methods
An outbreak of MDR Pa occurred on our vascular surgical ward during a 13-month period. Bacteria cultures positive for MDR Pa were obtained from 129 patients, and 64 CLI patients treated with IBS formed the study group. A control group of 64 was retrospectively matched from MDR Pa-negative patients treated with IBS in the same unit according to sex, age, presence of diabetes, Fontaine class, graft material, and site of the distal anastomosis. The most frequent sites of initial positive MDR Pa culture were the incisional wound in 30 (47%) and ischemic ulcer in 23 (36%). Median time between the positive MDR Pa-culture and IBS was 14 days (range, 56 days pre-IBS to 246 days post-IBS). Graft patency, survival, leg salvage, and amputation-free survival were assessed.
Results
One-year amputation-free survival (± standard error) was 52% ± 6% in the MDR Pa group vs 75% ± 5% in the control group (P = .02). Five-year amputation-free survival was 29% ± 6% in the MDR Pa group and 32% ± 6% in the control group (P = .144). For MDR Pa and control groups, the 1-year survival was 69% ± 6% and 82% ± 5% (P = .063), respectively, and 5-year survival was 36% ± 6% and 36% ± 6% (P = .302), respectively. For the MDR Pa and control groups, leg salvage was 79% ± 5% and 92% ± 4% at 1 year (P = .078) and 73% ± 7% and 87% ± 5% at 5 years (P = .126), respectively. The overall secondary patency rate at 1 year was 72% ± 7% in the MDR Pa group vs 81% ± 6% in the control group (P = .149). Local wound surgery was more frequent in MDR Pa patients than in controls (P = .002).
Conclusions
The MDR Pa outbreak was associated with a decreased short-term amputation-free survival after IBS for CLI in patients with positive MDR Pa culture. The potential risks of MDR Pa should be seriously considered whenever a positive culture is obtained in a vascular patient with CLI.
Pseudomonas aeruginosa (P aeruginosa) is a leading cause of hospital-acquired infections.1 Patients hospitalized for a long period are frequently colonized by P aeruginosa and are at increased risk of developing infection.2 The prominence of the P aeruginosa pathogen is related to its ability to develop resistance against almost all antibiotics and many disinfectants.3, 4 The clinical relevance of multidrug resistant P aeruginosa (MDR Pa) has been stressed by reports of nosocomial infections and outbreaks among critically ill patients and in intensive care units in several countries.2, 5, 6, 7, 8
Antimicrobial resistant pathogens are increasingly involved in vascular surgical site infections.9 Wound and graft infections are potentially life- and limb-threatening complications in vascular surgery. The knowledge about consequences of a MDR Pa outbreak among patients with critical leg ischemia (CLI) is sparse. The primary aim of this study was to evaluate amputation-free survival rates after an outbreak of MDR Pa on our vascular surgical ward among the patients treated with infrainguinal bypass surgery (IBS) for CLI. The secondary aim was to assess graft patency, survival, leg salvage, and reintervention rates.
Patients and methods
The study protocol was approved by the Institutional Review Board of Helsinki University Central Hospital (Department of Surgery) in 2006.
Definition and description of the MDR Pa-outbreak
In April 2000, the first positive MDR Pa culture was obtained and recognized from a patient treated on our vascular surgical ward in Helsinki University Central Hospital. The strain of P aeruginosa was resistant to ciprofloxacin, tobramycin, and a combination of piperacillin and tazobactam. As determined from the first 50 samples, the same P aeruginosa clone caused all but one of the MDR Pa contaminations.
After the outbreak was recognized, repeated bacterial specimens were taken from chronic ulcerations and surgical wounds. Grafts, urine, or blood were cultured as required. Grouping of the patients according to MDR Pa status and the presence of chronic ulcers was initiated in August 2000. An infectious diseases consultant evaluated every patient with a positive MDR Pa result. Patients with signs of MDR Pa infection, including erythema and swelling around the tissue lesion, infection of prosthetic material, or positive blood culture, were treated with MDR Pa-directed microbial therapy. MDR Pa-directed microbial therapy was not administered to patients whose cultures from a superficial site were positive for MDR Pa but were without signs of infection and asymptomatic patients with a positive urine culture.
Eventually in March 2001, positive samples were obtained from biofilms in the showerhead in one of the washrooms. The same showerhead had been cultured negative 3 months earlier. The showerhead was replaced, and the outbreak was controlled within 1 month. Between April 2000 and May 2001, 129 patients at our vascular unit presented with MDR Pa, which represented 15% of the annual number of in-hospital patients.
Patient evaluation
The study comprised the 64 patients who had a positive MDR Pa culture and were treated with IBS for CLI during the outbreak at our vascular ward. The hospital vascular registry (HUSVasc) data of 517 patients treated with IBS for CLI between January 2000 and June 2003 were scrutinized to identify controls for the MDR Pa-positive patients. A control patient with negative MDR Pa culture was matched to each study patient according to sex, age ± 10 years, presence of diabetes, Fontaine class III or IV,10 vein or prosthetic graft material, and site of the distal anastomosis (proximal popliteal artery, distal popliteal artery or crural/pedal artery; Table I). The medical records of the cases and controls were re-evaluated in 2008. Survival data were gathered from the Population Register Centre. Follow-up for this study terminated 5 years after IBS.
Table I. Six factors were used for matching 64 patients infected with multidrug-resistant Pseudomonas aeruginosa with control patients
| Factor |
|---|
| Sex |
| Age ± 10 years |
| Presence of diabetes mellitus |
| Fontaine class |
 •III (rest pain), or |
| •IV(ischemic tissue lesion) |
| Graft material |
| •vein, or |
| •prosthesis |
| Site of the distal anastomosis |
| •proximal popliteal artery, or |
| •distal popliteal artery, or |
| •crural/pedal artery |
Study group demographics
Median age of the 64 patients was 75 years (range, 37-91 years). There were 34 men (53%) and 30 women, and 36 patients (56%) had diabetes. The indication for the IBS was an ischemic ulcer (Fontaine IV) in 55 legs (86%). Nine patients had rest pain (Fontaine III). Median preoperative ankle-brachial index (ABI) was 0.41 (range, 0.26-2.7). The site of the distal anastomosis of the infrainguinal bypass was the proximal popliteal artery in 3 legs (5%), the distal popliteal artery in 14 (22%), and the crural or pedal artery in 48 (73%). Vein grafts were used in 56 reconstructions (87%), and prosthetic grafts were used in 8 (Table II).
Table II. Clinical data of the 64 patients with multidrug-resistant Pseudomonas aeruginosa and their controls
| Demographics | Cases | Controls | P |
|---|---|---|---|
| Age, median (range) | 75(37-91) | 75(41-91) | .947a |
| ABI, median (range) | 0.41(0.26-2.70) | 0.37(0.21-1.84) | .339 |
| BMI, median kg/m2 (range) | 23(19-44) | 24(16-35) | .417 |
| Cerebrovascular disease, No. (%) | 9(14) | 8(13) | .795 |
| Coronary heart disease, No.(%) | 34(53) | 44(69) | .07 |
| Diabetes mellitus, No.(%) | 36(56) | 36(56) | >.99a |
| Distal anastomosis, No.(%) | |||
| Proximal popliteal artery | 3(5) | 3(5) | >.99a |
| Distal popliteal artery | 14(22) | 14(22) | >.99a |
| Crural/pedal artery | 47(73) | 47(73) | >.99a |
| Fontaine class, No.(%) | |||
| III | 9(14) | 9(14) | >.99a |
| IV | 55(86) | 55(86) | >.99a |
| Graft material, No.(%) | |||
| Vein | 56(88) | 56(88) | >.99a |
| Prosthesis | 8(12) | 8(12) | >.99a |
| Pulmonary disease | 12(19) | 15(23) | .516 |
| Serum creatinine mg/dL, median (range) | 97(40-626) | 96(53-195) | .775 |
| Sex, No.(%) | |||
| Male | 35(54) | 35(54) | >.99a |
| Female | 30(46) | 30(46) | >.99a |
aIndicates that the patients are matched according to the factor. |
P aeruginosa culture
The initial positive MDR Pa culture was obtained from the incisional wound in 30 patients (47%), from the ischemic ulcer in 23 (36%), from wounds with undefined location in 10 (16%) and from the urine in 1 (1%). The MDR Pa culture was positive before IBS in 10 patients (15%) and after IBS in 55 (85%). Median time between a positive MDR Pa-culture and IBS was 14 days post-IBS (range 56 days pre-IBS to 246 days post- IBS). Intravenous MDR Pa-directed microbial therapy was used in 43 patients (67%) with signs of MDR Pa-infection. A drug-sensitive strain of P aeruginosa was cultured from the ischemic ulcers or incisional wounds in 31 control patients (48%).
Antimicrobial therapy
In patients with chronic ulcers, the antimicrobial treatment was selected according to the bacterial culture and continued until the wounds had healed. The corresponding microbial therapy was given intravenously during surgery. If a patient did not have chronic ulcers, cefuroxime was administered for surgical prophylaxis. If a prosthetic graft was used, patients also received vancomycin for surgical prophylaxis. Patients with infections caused by MDR Pa were treated with MDR Pa-directed intravenous microbial therapy at least for 2 weeks, if not contraindicated. A typical regimen was meropenem combined with amikacin or ceftazidime.
Statistical analysis
For the statistical analysis, we used SPSS 16.0 software (SPSS, Chicago, Ill). Differences between groups were evaluated with the χ2 and Mann-Whitney U test. Kaplan-Meier method was performed for comparison of variables between patient groups over time (univariate analysis). Mean data are presented with the standard error. A value of P ≤ .05 was considered statistically significant.
Results
Amputation-free survival
The 1-year amputation-free survival was 52% ± 6% in the MDR Pa group compared with 75% ± 5% in the control group (Fig 1). The amputation-free survival was significantly decreased in patients with MDR Pa at 1 year after IBS compared with the control group (P = .02, χ2 test). The amputation-free survival rate was 29% ± 6% at 5 years in the MDR Pa group and 32% ± 6% in the control group. Kaplan-Meier analysis did not show a significant difference between the groups in the amputation-free survival during the 5-year follow-up (P = .144).
Fig 1.
Kaplan-Meier curves demonstrate amputation-free survival in the multidrug-resistant Pseudomonas aeruginosa (MDR Pa) group and in the control group. At 1 year after the bypass, patients with MDR Pa showed significantly decreased amputation-free survival (P = .020; χ2 test). After 5-year follow-up, the amputation-free survival was poor in both groups (P = .144, Kaplan-Meier analysis). The standard error is <10% at each time point in the curve.
Survival
The survival was 69% ± 6% at 1 year and 36% ± 6% at 5 years in the MDR Pa group compared with 82% ± 5% and 36% ± 6% in the control group (Fig 2). Differences in the survival between the MDR Pa group and the control group were not significant at 1 year (P = .063, χ2 test) or after the 5-year follow-up (P = .302, Kaplan-Meier analysis).
Fig 2.
Survival rates after infrainguinal bypass surgery due to critical limb ischemia in the multidrug-resistant Pseudomonas aeruginosa (MDR Pa) group and in the control group demonstrated by Kaplan-Meier analysis. The survival was lower in the MDR Pa group at 1 year compared with the control group, but the difference did not reach statistical significance (P = .063). Both groups showed low survival rates in long-term follow-up (P = .302). The standard error is <10% at each time point in the curve.
Leg salvage
The 1-year leg salvage was 79% ± 5% in the MDR Pa group compared with 92% ± 4% in the control group (P = .078. χ2 test). The leg salvage was 73% ± 7% at 5 years in the MDR Pa group and 87% ± 5% in the control group (P = .126; Kaplan-Meier analysis; Fig 3).
Fig 3.
Kaplan-Meier curves demonstrate the leg salvage. The leg salvage rate was lower in the multidrug-resistant Pseudomonas aeruginosa (MDR Pa) group at 1 year (P = .078) and at 5 years after infrainguinal bypass surgery (P = .126) compared with the control group, but the differences were not significant. The standard error is <10% at each time point in the curve.
Patency
The 1-year primary patency, assisted-primary patency, and secondary patency rates were 60% ± 7%, 67% ± 7%, and 72% ± 7% for the MDR Pa group and 64% ± 8%, 77% ± 7%, and 81% ± 6% for the control group (Fig 4). Univariate analysis and log-rank tests did not show any significant differences between the MDR Pa and control group in the primary (P = .222) and secondary patency rates (P = .149), but the assisted-primary patency rate was worse in the MDR Pa group (P = .049).
Fig 4.
Kaplan-Meier life-table curves demonstrate secondary patency in the multidrug-resistant Pseudomonas aeruginosa (MDR Pa) group and in the control group. The difference in the secondary patency rate between the MDR Pa group and the control group was not significant (P = .149). Graft patency was monitored until 1 year after the infrainguinal bypass surgery, determined by clinical examination, duplex Doppler imaging, and the ankle-brachial index. The standard error is <10% at each time point in the curve.
Vascular reintervention rates
Vascular reinterventions were required in 17 patients (27%) in the MDR Pa group and in 19 (30%) in the control group (Table III). Five of the eight prostheses (63%) were infected by MDR Pa and were removed. No prosthesis was infected or removed in the control group. The difference between the number of patients with vascular reinterventions in the MDR Pa group and the control group was not statistically significant (P = .694).
Table III. Vascular reinterventions performed in the multidrug-resistant Pseudomonas aeruginosa group and in the controls
| Vascular reintervention | Cases no. | Controls no. |
|---|---|---|
| Number of patients having vascular re-intervention | 17 | 19 |
| Vascular reinterventions, total | 29 | 28 |
| Distal extension/patch angioplasty | 7 | 8 |
| Suturation | 6 | 3 |
| PTA | 1 | 7 |
| Thrombectomy | 3 | 5 |
| Thrombolysis | 3 | 4 |
| Removal of prosthesis | 5 | 0 |
| Redo bypass | 4 | 1 |
Local wound surgery
In the MDR Pa group, 40 patients (63%) had local surgery, as did 21 patients (33%) in the control group (Table III). Significantly more patients in the MDR Pa group had local ulcer surgery than those in the control group (P = .002; Table IV).
Table IV. Local wound surgery and major amputations performed in the multidrug-resistant Pseudomonas aeruginosa group and in the controls
| Wound surgery | Cases no. | Controls no. |
|---|---|---|
| Total patients needing local wound surgery | 40 | 21 |
| Total number of local wound surgery | 49 | 25 |
| Minor amputationa | 26 | 17 |
| Revision | 18 | 5 |
| Skin grafting | 3 | 2 |
| Microvascular flap | 2 | 1 |
| Major amputations | 14 | 8 |
aTransmetatarsal level or distal to the transmetatarsal level. |
Discussion
Infrainguinal revascularizations are useless unless life and limb are preserved. The considerably decreased amputation-free survival at 1 year in the MDR Pa group compared with the control group suggests that MDR Pa in a patient with CLI should always be considered a clinical challenge for the patient's outcome. The control patients were meticulously matched to the MDR Pa patients in our study. The most vulnerable patients with MDR Pa died early after IBS or were apt to have a major amputation soon after the reconstruction due to nonhealing ischemic ulcers or incisional wounds. Later on during the follow-up, the differences in the amputation-free survival rates between the MDR Pa and the control group disappeared. The explanation behind the lack of influence of the MDR Pa on long-term follow-up may be that the overall long-term survival of patients with CLI is low. The survival rate was 36% at 5 years in the control group. Furthermore, owing to the small number of patients, our study might lack power to adequately analyze the differences between subgroups.
Resistance to antimicrobial agents is an increasing public health threat.1 To our knowledge, this is the first study analyzing the consequences of a MDR Pa outbreak among patients with CLI. Six other case-control studies have associated MDR Pa with increased death.10 The in-hospital mortality rates among patients with MDR Pa varied between 21% and 54% in those studies.1, 6, 7, 8, 11 These high in-hospital mortality rates is be explained by the patient series, which consisted mostly of patients treated in intensive care units. Multimorbid patients cannot be compared directly with the patients in our study, although our CLI patients had several comorbidities. The patient cohort in our study was unique, because all patients had well-defined disease and a surgical procedure.
One study did not find any effect of MDR Pa on death,12 but the bacteria in that study were already considered MDR if they were resistant to two or more antipseudomonal antibiotics. Because there is no international consensus on the definition of MDR in P aeruginosa, the results in different studies cannot be directly compared.
The site where the MDR Pa is isolated from may influence short-term mortality. Most of the initial MDR Pa isolates in this study were from incisional wounds and ischemic ulcers, whereas the initial MDR Pa in other studies has frequently been a respiratory isolate.6, 7 Bloodstream infection caused by MDR Pa causes considerable death.13
Because in almost half of the control group a non-MDR Pa was cultured from the ischemic ulcers or incisional wounds, our study indicates that the MDR character of P aeruginosa has a crucial role for the decreased short-term rates of amputation-free survival. The initial MDR Pa infection was in the incisional wound in one-third of the study patients. The incisional wound infection rate has been reported to be between 2% and 54%, depending on the type of surgery, patient demographics, and definition of wound infection.9, 14, 15, 16 In this study, we did not analyze wound infection caused by other bacteria. The limited available means of effectively treating MDR Pa led us to perform more local wound surgery among the study patients to eradicate the source of bacteria. Local wound care could not be analyzed in this retrospective study. The positive MDR Pa result might have influenced the treatment strategy in some patients.
Although few patients in our study had prosthetic grafts, we want to stress the poor outcome of these patients in the MDR Pa group. Five of eight prostheses were infected in the MDR Pa group; two patients whose prostheses were removed had a new bypass with vein and avoided major amputation, whereas the three nonrevascularized patients underwent a major amputation during the follow-up. No infection of the prosthetic grafts occurred in the control group. The reported incidence of vascular prosthetic graft infections varies between 0.2% and 29%, depending on the location of the graft, indication for intervention, and host defence mechanisms.9, 17
The greatest problem during the outbreak was identifying the source of the MDR Pa. Despite several attempts to track the origin, it remained obscure. In addition to the clinical consequences, the outbreak caused additional hospital burden, including cohorting, hygiene control efforts, bacterial culture sampling, microbial therapy, and increased hospital length of stay. The economic effect of the outbreak was not analyzed in this study.
Conclusion
MDR Pa was associated with decreased rates of short-term amputation-free survival in patients treated with IBS due to CLI compared with a matched cohort of non-MDR Pa patients. In addition, excess local surgical interventions for wound treatment and prosthetic graft removals were required due to MDR Pa. MDR Pa in a patient with CLI should always be considered as a serious event with high risk of adverse clinical consequences.
Author contributions
References
- . Multi-drug resistant Pseudomonas aeruginosa; risk factors and clinical impact. Antimicrob Agents Chemother. 2006;50:43–48
- . Pseudomonas aeruginosa carriage, colonization, and infection in ICU patients. Intensive Care Med. 2007;33:1155–1161
- Pseudomonas aeruginosa may accumulate drug resistance mechanisms without losing its ability to cause bloodstream infections. Antimicrob Agents Chemother. 2007;10:3531–3536
- . Resistance mechanisms in Pseudomonas aeruginosa and other nonfermentative gram-negative bacteria. Clin Infect Dis. 1998;27(suppl 1):S93–S99
- Outbreak caused by Tobramycin-resistant Pseudomonas auruginosa in a Bone Marrow Transplantation Unit. Scand J Infect Dis. 2001;33:445–449
- . Risk factors and clinical outcomes of nosocomial multi-drug resistant Pseudomonas aeruginosa infections. J Hosp Infect. 2004;57:112–118
- The influence of metallo-ß-lactamase production on mortality in nosocomial Pseudomonas aeruginosa infections. J Antimicrob Chemother. 2006;58:387–392
- . Fluoroquinolone-resistant Pseudomonas aeruginosa assessment of risk factors and clinical impact. Am J Med. 2006;119:526e19-25
- . vascular surgical site infection: risk factors and preventive measures. Semin Vasc Surg. 2008;21:119–123
- . Die chirurgische Behandlungen der peripheren Durchblutungsstörungen. Helv Chir Acta. 1955;36:499–533
- . Clinical and economic impact of common multidrug-resistant gram-negative bacilli. Antimicrob Agents Chemother. 2008;52:813–820
- . Endemic multi-drug resistant Pseudomonas aeruginosa in critically ill patients. Infect Control Hosp Epidemiol. 2004;25:825–831
- Risk factors for antimikrobial resistance and influence of resistance on mortality in patients with bloodstream infections caused by Pseudomonas aeruginosa. Microb Drug Resist. 2005;11:68–74
- . Wound infections. Surg Clin North Am. 1994;74:519–536
- . Prospective study of wound complications in continuous infrainguinal incisions after lower limb arterial reconstructions: Incidence, risk factors and cost. Surgery. 1996;119:378–383
- Analysis of risk factors for surgical wound infections following vascular surgery. Am J Med. 1991;91:170–172S
- . Infection in prosthetic vascular grafts. In: Rutherford editors. Vascular surgery. Philadelphia: Elsevier Saunders; 2005;p. 875
Competition of interest: none.
PII: S0741-5214(09)01228-2
doi:10.1016/j.jvs.2009.05.063
© 2009 Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
