Sepsis caused by multidrug-resistant klebsiella pneumoniae infection in a 23-year-old burn patient: case report and literature review

Summary

Septic complications are a major problem in burn care. Infections caused by multidrug-resistant (MDR) Klebsiella pneumoniae strains represent a growing threat. In particular, Romania has seen rapidly increasing rates of MDR Klebsiella pneumoniae isolates in recent years. The PubMed database was searched with the terms ‘Klebsiella pneumoniae’, ‘multiple drug resistant’, ‘enterobacteria’, ‘infection’, ‘sepsis or septicaemia or blood stream infection’, ‘surveillance’, ‘diagnose’, ‘burn’, ‘intensive care unit’ and ‘Romania or Europe’. In addition, the case of a burn patient who suffered from sepsis caused by MDR Klebsiella pneumoniae and who was treated at our burn intensive care unit (ICU) was reviewed. The 23-year-old male was initially treated in Romania, but after 48 hours he was transferred to our burn ICU. Initially, the treatment course was appropriate and sufficient wound healing was achieved by day 27. However, the patient’s condition deteriorated rapidly on day 28 and blood culture analysis showed growth of MDR Klebsiella pneumoniae. Despite intensive medical therapy, the patient died as a result of multiple organ failure (MOF) on day 44. Sepsis caused by MDR Klebsiella pneumoniae represents an enormous therapeutic challenge. Burn patients who are transferred from Southeast Europe would likely benefit from an intensified screening for MDR pathogens.

Introduction

Severely burned patients requiring intensive care are immune-compromised and are therefore at a high risk of developing septic shock and MOF.¹ One reason for this is the loss of the barrier function of the skin.² Moreover, the use of intravascular devices raises the risk for serious complications, such as thrombophlebitis and catheter-related bloodstream infections.³ It has been shown that infections caused by MDR pathogens are a negative prognostic sign. Infections in burn patients caused by MDR bacteria are associated with prolonged ICU stays and higher mortality rates.^4-6 The European Centre for Disease Prevention and Control (ECDC) reports a significant increase in the prevalence of multidrug-resistant Gram-negative bacteria strains in Southeast Europe.⁷ In 2015, Romania had the third highest rate (49.8%) of cultured MDR Klebsiella pneumoniae strains in Europe.⁸ In contrast to Romania, the rate of MDR Klebsiella pneumoniae in Austria was relatively low in 2015 (3.3%).⁸

Materials and methods

The PubMed database was searched with the terms ‘Klebsiella pneumoniae’, ‘multiple drug resistant’, ‘enterobacteria’, ‘infection’, ‘sepsis or septicaemia or blood stream infection’, ‘surveillance’, ‘diagnose’, ‘burn’, ‘intensive care unit’ and ‘Romania or Europe’. All results of the literature search were reviewed and relevant information was extracted, summarized and discussed.

In addition, we conducted a retrospective medical chart review of a burn patient who suffered from sepsis caused by MDR Klebsiella pneumoniae and who was treated for burn trauma at the ICU of the General Hospital of Vienna.

Case report

We report a case of a 23-year-old male who was injured after playing with flammable materials. The patient did not have any known prior medical conditions. He suffered from deep dermal and full thickness burns of the face, trunk and both arms. The Total Body Surface Area (TBSA) of the burns was estimated at 42% and was determined by the Wallace rule of nines (Fig. 1A-E). The probability of survival, calculated with the Abbreviated Burn Severity Index (ABSI), was 50-70% (ABSI: 8).⁹ The patient was admitted to a Romanian hospital immediately after his burn injury. It took 48 hours to coordinate the transfer of the patient to our burn ICU by airplane and emergency ambulance. At the time of arrival, the patient was hemodynamically stable, intubated and mechanically ventilated. The day of admission to our burn ICU was defined as day 0.

Fig. 1. Burn injury on day of admission (48 hours after the accident).

On day 0, a tracheostomy was performed. Microbiological screening on day 0 showed heavy colonization with multiple pathogens. The nose and throat region were heavily colonized by Methicillin-resistant Staphylococcus aureus (MRSA), Klebsiella pneumoniae, Hafnia alvei, Escherichia coli and Candida albicans. The urine culture was negative. The perianal region showed growth of Escherichia coli and Enterococcus faecalis. On day 2, wound swabs of the face and upper limbs were taken and showed heavy colonization with MRSA, Klebsiella pneumoniae, Hafnia alvei, Escherichia coli and Candida albicans. The initial antibiotic therapy of cefepime (2 x 2000 mg per day) followed by daptomycin (1 x 850 mg per day) was instituted and continued until day 11.

The surgical team was able to perform the first reconstructive surgery on day 3. A tangential necrosectomy of the ventral thorax and both arms was performed. Wound coverage was achieved using autologous split-skin grafts from the ventral thighs (1:3). After two weeks of treatment, the wounds showed sufficient healing (Fig. 2A-D). However, a further operation was necessary on day 17 to treat thoracic and axillary residual defects. Again, wound coverage was performed using autologous split-skin grafts from the ventral thighs (1:1.5). The facial burns were treated with Acticoat-7 and healed conservatively by day 17 (Fig. 1E and Fig. 2A). The patient was kept in a FluidAir bed, which enabled the wounds on his back to heal after 16 days without surgical intervention (Fig. 1D and Fig. 2D).

Fig. 2. Wound status after 16 days of treatment.

On day 10, wound swabs showed colonization with extended-spectrum β-lactamase (ESBL) and carbapenemase producing Klebsiella pneumoniae. Antibiotic therapy with meropenem (3 x 2000 mg per day) was administered from day 9 to 21. From day 19 to 21, fosfomycin (3 x 8000 mg per day) was included in the patient’s antimicrobial treatment.

On day 26, the patient was successfully decannulated, and the burn wounds showed improved healing. At that point, the patient was fully conscious, responsive and mobile. Furthermore, return transportation to his hometown had been arranged.

Daily blood tests did not reveal evidence of systemic inflammation from day 20 to 27. Between days 27 and 29, a dramatic increase in c-reactive protein (CRP) (6.03 mg/dL to 27.8 mg/dL) and leucocytes (14.96 G/L to 58.95 G/L) was observed. Starting on day 28, the patient developed a fever with a temperature as high as 39.9°C, tachycardia (from a mean baseline rate of 75 bpm to a mean of 110 bpm) and his mean arterial blood pressure (MAP) decreased significantly (from a mean of 110 mmHg to a mean of 50 mmHg). Empiric antimicrobial therapy with ceftolozane/tazobactam (3 x 3000 mg per day) was started immediately. Since we suspected a bloodstream infection, the central venous catheter (CVC) was removed on day 29. Shortly after removal, the patient developed ventricular fibrillation and had to be intubated and revived by cardiopulmonary resuscitation for several hours. During the resuscitation, an extracorporeal membrane oxygenation (ECMO) system was implanted.

The microbial cultures of the peripheral and central blood samples as well as the CVC-tip and the rectal swab showed heavy growth of EBSL and carbapenemase- (class D - OXA-48) producing Klebsiella pneumoniae. Furthermore, on day 29, the antibiograms of these blood cultures showed antimicrobial susceptibility to only fosfomycin and tigecycline. However, the Minimal Inhibitory Concentrations (MICs) for fosfomycin and tigecycline were not specified at that time. Based on this information, the antibiotic therapy was changed to fosfomycin (3 x 8000 mg per day) and tigecycline (1 x 300 mg per day) on day 30.

To improve oxygenation, the patient needed continuous support by ECMO until day 37. In addition, renal replacement therapy (citrate dialysate) was necessary (initial value of plasma creatinine 2.18 mg/dL, diuresis 300 mL/24h on day 29). For hemodynamic stabilization, epinephrine (up to 0.196 μg/kg/min) and noradrenaline (up to 0.549 μg/hg/min) were required. The patient developed Disseminated Intravascular Coagulation (DIC) and needed transfusions of large quantities of erythrocytes, platelets and fresh frozen plasma. A heparin-induced thrombocytopenia (HIT) was immunologically excluded. A computed tomography scan of the thorax and abdomen revealed thrombosis in the left brachiocephalic vein and in the right femoral vein (reaching the inferior vena cava), of which both were former CVC locations.

Over the course of treatment, the patient’s status progressively worsened, and on day 40 the antibiotic treatment was changed to trimethoprim/sulfametrole (3 x 500 mg per day). Due to ischemia of the small intestines, an ileostomy was performed on day 42. Throughout the therapeutic process, we were unable to eradicate the MDR Klebsiella pneumoniae.

As a result of the progressive deterioration of the patient´s overall condition (lactate >14 mmol/L, MAP > 50 mmHg under 1 μg/kg/min noradrenaline), the young patient succumbed to his infection-associated MOF on day 44.

Discussion

Distribution of multidrug-resistant Klebsiella pneumoniae A vast increase in the spread of infections caused by MDR Klebsiella pneumoniae strains has been reported worldwide.¹⁰ In particular, the prevalence of isolated MDR Klebsiella pneumoniae strains in Romania has increased rapidly from a rate of 11.8% in 2009 to a rate of 49.8% in 2015, as reported by the ECDC.^7,8 In contrast, the rate of MDR Klebsiella pneumoniae increased from 2.0% to 3.3% in Austria during the same period.⁸

In our retrospective case study, the probability is nearly 50% that the Klebsiella pneumoniae that first colonized the patient’s wounds was resistant to third-generation cephalosporins, fluoroquinolones and aminoglycosides. Two models are offered to explain why we did not detect the presence of an MDR organism until after day 10: 1) at the time of admission, the patient had various Klebsiella pneumoniae clones with dissimilar microbial loads on several wounds and body regions; and 2) the antibiotic therapy that was initially used was selected for MDR clones of Klebsiella pneumonia.¹¹

Approximately two-thirds of Klebsiella septicaemias are caused by multidrug-resistant strains.¹² Recent analyses have shown increased resistance to penicillin, third generation cephalosporins, fluoroquinolones and aminoglycosides mediated by the production of ESBL, AmpC β-lactamase and carbapenemase for Klebsiella pneumoniae strains in Europe.^12-14

Regardless of the cause of infection or the comorbidities, blood stream infections caused by antibiotic-resistant Klebsiella pneumoniae are associated with a mortality rate of 24% to 72%.^15,16

Recently, resistance to reserve antibiotics such as polymyxin has been reported. This suggests a dramatically increased probability of the emergence of panresistant bacterial clones, which will soon further complicate the likelihood of successful therapy for Gram-negative infections.^7,13

Antibiotic treatment

To guarantee high rates of survival in critically ill patients suffering from bacterial bloodstream infections, prompt institution of appropriate empiric antimicrobial therapy in the early phase of septicaemia is important.¹⁷ In particular, for septicaemias caused by ESBL-producing enterobacteria, delayed administration of appropriate antibiotic therapy was observed five times more often than in non-antibiotic resistant enterobacteria septicaemias, resulting in significantly (two-fold) higher mortality rates for the antibiotic resistant septicaemias.18 In 2014, Girometti et al. concluded that only 23% and 33%, respectively, of patients with ESBL or carbapenemase producing Klebsiella pneumoniae septicaemias received adequate empirical antibiotic therapy (versus 74% in non-resistant Klebsiella pneumoniae septicaemias).¹²

In the present case, appropriate antimicrobial therapy with fosfomycin and tigecycline was instituted on day 30 (two days after first symptoms) when microbial culture confirmed a CVCassociated septicaemia with MDR Klebsiella pneumoniae. In addition, the antimicrobial susceptibility of the isolated pathogen was identified only semi-quantitatively, and the MIC was not specified either for fosfomycin or for tigecycline.¹⁹ Therefore, both antibiotics were dosed empirically. On the basis of the patient´s country of origin, injury and clinical course, the empiric antimicrobial therapy (ceftolozane/tazobactam), which was started on day 28, should have been considered earlier in light of the suspicion of a probable multidrug-resistant Klebsiella pneumoniae-associated septicaemia; this would have led to earlier aggressive antimicrobial therapy while waiting for the antibiotic susceptibility to be ascertained from the culture results. However, the isolated MDR Klebsiella pneumoniae strain from the rectal and wound swabs was found to be susceptible to ceftolozane/tazobactam with an MIC of 16 mg/l for the rectal sample and an MIC of 32 mg/l for the wound sample. At the same time, blood cultures demonstrated no susceptibility to these antibiotics.

The current literature discusses several recommendations for treating infections caused by MDR Gram-negative bacteria, especially for carbapenemase producing Klebsiella pneumoniae. Combination antibiotic therapy consisting of first and second line antibiotics has shown better outcomes for patients than has monotherapy.¹² In comparison with other protocols, combination antibiotic therapy consisting of carbapenem or fosfomycin (alone or together) along with tigecycline, colistin or an aminoglycos, was shown to have the lowest mortality rates for infections caused by MDR Gramnegative bacteria, as well as for carbapenemase-producing bacteria.^20-24

Additional diagnostic strategies

One factor in the high rate of delayed or inappropriate antimicrobial therapy in patients suffering from antibiotic-resistant septicaemias may be explained by the time it takes to complete the microbial analysis. Commonly used (phenotypic) blood cultures take 12 h to several days to deliver reliable results in the identification of the organism and its antimicrobial susceptibility.²⁵ Furthermore, if isolates show complex resistance mechanisms (i.e., more than one beta-lactamase such as ESBL and AmpC beta-lactamases), phenotypic tests may be insufficient. ^26,27 For this reason, as well as the need to minimize the length of time needed for a microbial diagnosis especially in critically ill patients, the combination of genotypic methods (e.g., polymerase chain reaction, nucleic acid isothermal amplification, and next-generation sequencing) and conventionally used phenotypic tests is recommended by some authors and may optimize diagnosis and improve patient outcomes.^12,25,27,29 In the present case, microbial testing followed EUCAST (European Committee on Antimicrobial Susceptibility Testing) guidelines and involved both phenotypic and genotypic methods.³⁰ Another factor in the delay of diagnosis of concurrent septicaemia in ICU burn patients is the Systemic Inflammatory Response Syndrome-like state that is caused by severe burn trauma and which often masks the beginning of septicaemia. Various studies have shown that the use of biomarkers such as serum concentrations of procalcitonin (PCT) or interleukin-6 (IL-6) can shorten the diagnostic window for septicaemias, thereby increasing the rate of expeditious initiation of empiric or appropriate antibiotic treatment.^2,3 For this reason, a daily screening of the patient´s PCT-serum concentration is recommended.³

Regular scans of CVC-bearing veins to detect septic thrombophlebitis and thrombotic deposits can help direct the early removal of CVCs and has been shown to be a highly significant factor in decreasing blood-stream infection-related mortality in neutropenic patients.³² This strategy may also lower CVC infection-related mortality in severely burned patients.³³

Finally, prognostic scores such as the ABSI score do not account for the presence of MDR pathogens or the patient´s country-specific MDR distribution. In the present case, the patient´s ABSI score yielded a midrange score of 8 at the time of admission (corresponding to a probability of death between 30-50%). In recent years, we have treated other patients from Southeast Europe whose clinical courses were comparable to that of the present case, and we believe that this scoring system underestimates the probability of death for patients with MDR infections at the time of admission and for those who have been previously treated in countries with high MDR prevalence. Reevaluation and potential extensions of generally used prognostic scoring systems may be indicated to update and enhance their usefulness with regard to MDR infections and countries of origin.

Our case highlights the high risk of developing serious and difficult-to-treat infections caused by MDR organisms, especially in patients transferred from Southeast Europe. The combined use of continuous extended screening and analysis methods for diagnosing septicaemias and the implementation of protocols to promote earlier replacement of CVCs may be indicated to improve outcomes in this patient group. We will continue to investigate the usefulness of these strategies for treating internationally transferred patients in our burn ICU.

Conclusion

Steady improvements in acute treatment, intensive care and surgical interventions have led to a clear decrease in burn-associated mortality in the past 30 years.³⁴ However, infections caused by MDR pathogens increase the risk of fatality in burn patients. Currently, septicaemia and MOF are the most common causes of death for severely burned patients, as has been described in the present case and literature review.^1,35 Consequently, a microbial screening (wound swab, throat and nasal swabs, and a rectal or perirectal swab) should be performed routinely at the time of patient admission.

Moreover, prognostic scores such as the ABSI do not take into consideration the presence of MDR pathogens or the patient’s individual risk for developing infections caused by MDR pathogens. This may lead to an underestimation of the probability of death.⁹ In particular, burn patients requiring intensive care who are transferred from Southeast Europe may benefit from intensified screening for MDR bacteria.