Our objectives today are to understand the etiologies and clinical manifestations of life-threatening infections, to choose appropriate antimicrobials for patients with life-threatening infection, and to understand the mechanism of action, spectrum of activity, and potential adverse effects of various antimicrobials. We'll start with a case. This is a case of a seven-month-old unimmunized female admitted with a history of fever, emesis, and irritability. Exam findings are significant for a fever, a tense, bulging anterior fopnal, a stiff neck, evidence of sepsis with tachypnea and tachycardia, and an irritable infant. Lab findings are consistent with inflammation, including leukocytosis, elevated ESR and CRP, as well as mild hyponatremia. Other studies are pending. The chest X-ray is normal. A SCT is concerning for mild dilation of the ventricles. The questions are, what is the most likely infection and organism, and what empiric antibiotics will you start? Now, we're going to come back to this case toward the end. First, we're going to talk a little bit about why antibiotic strategies are so important in the PICU. It turns out that antibiotics are some of the most commonly prescribed medicines in the PICU population, with up to 80% of PICU patients receiving at least one antimicrobial agent during their stay. And it's been estimated that up to about a half of antibiotics could be considered inappropriate due to either too broad or too narrow a spectrum of coverage or inappropriate dosing. When we consider antibiotic selection in the PICU, we really need to consider four different elements. First, what organism or organisms do we suspect is causing the infection? Second, what body site is involved? And third, what is the ability of the antibiotic to actually get into that body site? And fourth, of course, is the susceptibility of the organisms that we suspect to the antibiotics that we are choosing. Considering the etiologies of life-threatening infection in the PICU, we can think about the etiologies of sepsis in critically ill children. These data are from a multicenter study of 795 children with community-acquired sepsis admitted to European PICUs. And what you can see in the chart is that bloodstream and CNS infections were predominant, followed by pneumonia and other sources of infection. Now, it's important to note here that those patients in this study who had culture-negative infections were presumed to have bloodstream infection. Looking at another study, this time of children with severe sepsis or septic shock admitted to 128 PICUs from around the world, we see here that respiratory infections were predominant, followed by bloodstream and then abdominal infections, with CNS infections only representing about 4% of those children, again, with severe sepsis or septic shock. Regarding causative organisms, we turn again to the same point prevalence study of 567 children with severe sepsis or septic shock admitted to PICUs internationally. In this study, about 35% of the patients had culture-negative sepsis. This is similar to other studies that have shown about a 30% to up to 50% rate of culture-negative sepsis in children. Among those with bacteria isolated, about an equal proportion had gram positive versus gram negative organisms, with Staph aureus and Pseudomonas species predominating. Fungal infections were present in about 13% of patients, with Candida species being the most frequent fungal infection identified. Viruses were present in about 21% of children, and these could be viral infections with or without concomitant bacterial infection. These studies can identify likely causative organisms on a population level, but when considering treatment for an individual patient in front of you, it's important to consider patient-specific factors as well. One of those factors is the age and immunization status of your patient. The other is geographic region and particularly local patterns of antimicrobial resistance, as well as the presence of chronic medical conditions and or frequent contact with the healthcare system. This includes diagnoses of immune compromise. These are important considerations in the PICU population because of the high prevalence of children with chronic medical conditions among those with severe infection in the PICU. As an example of age-related differences, we can consider differences in likely causative organisms for meningitis across the pediatric age range, with infants, particularly those up to three months of age, having a higher predominance of GBS, E. coli, and listeria infections compared to older infants and children, wherein strep pneumo and Neisseria meningitidis are predominant, with potentially hypophilus influenzae depending largely on immunization status. When considering the other factors, we're really trying to estimate the risk of a patient being infected with a multidrug-resistant organism. So let's review some of those multidrug-resistant organisms here, starting with methicillin-resistant Staphylococcus aureus. We know that community acquired MRSA rates vary widely depending on geographic location. And certainly if your local biome contains 10% or greater of Staph aureus species that are resistant to methicillin, then vancomycin really should be considered for empiric therapy, depending on the severity of the infection and consideration for individual risk factors for MRSA. Treatment options for MRSA generally include vancomycin or linazolid, as most MRSA species in the U.S., particularly in pediatrics, are susceptible to both vancomycin and linazolid. There are other options, as listed below, that remain alternatives as well. Drug resistance among Enterococcus species varies. And in general, at this point, vancomycin-resistant Enterococcus species are more prevalent in adult settings, though it's important to understand whether or not these organisms exist in your local biome or your hospital. Treatment options include a continuous infusion of ampicillin, or ampicillin plus a beta-lactamase inhibitor for penicillin-resistant Enterococcus species, or linazolid or daptamycin for vancomycin-resistant species. Multidrug-resistant gram-negative organisms are an emerging problem across multiple health care systems. These include extended-spectrum beta-lactamase-producing enterogram negatives, such as Klebsiella, Serratia, E. coli, and Enterobacter species. These extended-spectrum beta-lactamases render these bacteria resistant to penicillins. And notably, some Klebsiella species are now becoming resistant to carbapenems as well. Pseudomonas, Acinetobacter, and Stenotrophomonas are also important gram-negative species that can rapidly become multidrug-resistant in a population. When considering empiric therapy for these organisms, one really must consider the patient history, including a personal history of growth of a drug-resistant organism, as well as your local resistance patterns. It may be important to consider double-covering these bacteria until your sensitivities are available. Other options may include older antibiotics that had fallen out of favor due to side-effect profile, but are starting to regain use in selected circumstances of multidrug-resistant bacteria. As a theme, I think we can say that patterns of antimicrobial resistance vary widely, and so it's important to know your hospital's antibiogram. These antibiograms can be vitally important for identifying appropriate empiric coverage based on local resistance patterns. It's also important to consider patient-level risk factors for healthcare-associated organisms, which are often multiply drug-resistant. These are data from children admitted to a single PICU with at least one culture and treated with a course of antibiotics. The cohort was divided into 556 children as a derivation cohort and 525 children as a validation cohort, with about a quarter actually growing high-risk gram-negative rods. The sensitivity for both the derivation and validation cohort was excellent for the risk factors identified, with a moderate specificity. And what were those risk factors? They included being hospitalized for greater than 48 hours, recent antibiotic use, chronic lung disease, living in a chronic care facility, prior high-risk gram-negative rods, as well as recent hospitalization. And here, recent hospitalization was defined as hospitalization within the past one month. Children with immune compromise have specific infection risks, depending on their specific immune deficiency, with children with neutropenia or lymphopenia being susceptible to bacteria of all varieties, and those with asplenia or complement deficiencies being mostly susceptible to encapsulated organisms, including strep pneumoniae, haemophilus influenzae, and Neisseria meningitidis. Moving on to antibiotics, the slide here lists various classes of antibiotics for your reference. We'll get into these in more detail in the next few slides. The first family of cell wall inhibitors or beta-lactams we will consider are the penicillins. The penicillins comprise multiple subfamilies that each have different spectrums of activity. The immunopenicillins, or ampicillin, remain the drug of choice for enterococcal infections, though have some extended gram- negative coverage as well. Penicillinase-resistant or anti-staph penicillins, such as nafcillin and oxacillin, are the drugs of choice for MSSA infections, though they have no activity against MRSA. Extended-spectrum penicillins include beta-lactamase inhibitors. Some of these can be very useful for empiric treatment for severe healthcare-associated infection due to their anti-pseudomonas activity. Mechanism of resistance for penicillins include either beta-lactamase enzymes or altered transpeptidase binding sites. Adverse effects include anaphylaxis, rash, and rarely neutropenia or thrombocytopenia with prolonged use. Regarding cephalosporins, spectrum of activity also differs based on the generation. In general, when considering first through fourth generation cephalosporins, you have greatest gram-positive activity in the first generation that decreases through the generations as you increase gram- negative coverage. So your first-generation cephalosporins, including cefazolin, have excellent gram-positive coverage, though they don't cover MRSA or enterococcus. Second-generation cephalosporins, including cefoxitin, have better gram-negative and some anaerobic coverage, so they can be reasonable choices for enteric organisms. Your third- generation cephalosporins have fairly broad gram-positive and gram-negative coverage, and some have pseudomonas coverage as well. Fourth-generation cephalosporins, including cefepime, are similar to third-generation with the addition of penicillin-resistant pneumococcus, pseudomonas, and extended-spectrum beta-lactamase producing gram-negatives. The fifth-generation cephalosporins are similar to the third-generation, but with greater gram- positive coverage, including MRSA and penicillin-resistant pneumococcus. Vancomycin has an extensive gram-positive spectrum of activity, including MRSA and enterococcus. Oral vancomycin is the drug of choice for C. diff. colitis. Adverse effects include an anaphylactoid reaction, which involves puritis and an erythematous rash, particularly on the face, chest, and upper extremities. This reaction can be mitigated by slower infusion rates and antihistamine treatment. The more concerning adverse effect of vancomycin is renal toxicity, and because of the narrow therapeutic window of vancomycin, therapeutic drug monitoring really is essential. Carbapenems include the antibiotics you see listed there, though meropenem has largely replaced imopenem due to a greater safety profile and improved gram-negative coverage. Carbapenems have a broad spectrum of activity, though they are not effective against MRSA, stenotrophomonas, or certain enterococcus species. In general, the carbapenems really are reserved for severe infections with suspicion for drug resistance, and meropenem in particular is a frequent target of antibiotic stewardship programs. This is exceptionally important in order to try to decrease the incidence of carbapenem-resistant gram-negatives. Adverse effects of carbapenems include increased risk of seizures with imopenem, particularly in the setting of CNS infection. Otherwise, the carbapenems are generally pretty well tolerated. Colistin is an older antibiotic that had been removed from common use due to nephrotoxicity, though it's been suggested that this nephrotoxicity may have been due to differences in monitoring and dosing patterns in the past compared to the current era. Colistin's resurgence really has been for the treatment of severe healthcare-associated infections due to multi-drug-resistant gram-negatives. Colistin is available in inhaled or IV form, and it's important to note that colistin really has no role in anaerobic or gram-positive infections. Daptomycin works by disrupting the cell membrane. It has extensive anti-staphylococcal coverage, including MRSA, as well as anti-streptococcal coverage. It's generally observed as a second-line agent for severe gram-positive infections, and does have activity against both methicillin-resistant and vancomycin-resistant organisms. In terms of adverse effects, daptomycin can cause peripheral neuropathy as well as a myopathy, and in general, its routine use is not recommended in children under the age of 12. Regarding protein synthesis inhibitors, aminoglycosides have bactericidal activity against a wide range of bacteria, though they are most commonly used for gram-negative infections as well as some staph and enterococcal infections. There may be some synergy when combining aminoglycosides with a cell wall inhibitor due to the augmented uptake of aminoglycosides by susceptible bacteria, and it's important to note that there really is no role for aminoglycosides in CNS disease because of low penetration in the CSF. Adverse effects of aminoglycosides include nephrotoxicity and ototoxicity. Regarding nephrotoxicity, risks increase particularly with the use of other nephrotoxic drugs, and therapeutic drug monitoring is key. Linazolid has excellent gram-positive coverage and is often used for infections caused by resistant strep pneumo, MRSA, enterococcus, and vancomycin-resistant organisms. Linazolid also has excellent bioavailability into the tissues. Adverse effects include myelosuppression, a risk of serotonin syndrome, particularly in combination with other drugs, and linazolid can be associated with a duration-dependent lactic acidosis. This is thought to be due to decreased mitochondrial protein synthesis. Notably, unlike vancomycin, linazolid is not nephrotoxic and can sometimes be used in patients with concern for MRSA or other resistant gram-positive organisms in the setting of concern for renal insufficiency. Clindamycin has activity against many gram-positives and anaerobes, though is ineffective against enterococcus. Its most common use in the PICU is to inhibit protein synthesis and thereby toxin synthesis in the setting of toxic shock in gram-positive sepsis. Clindamycin can be used for community-acquired MRSA, though has high rate of resistance and can result in macrolide-inducible resistance, as demonstrated by the D-test here. Macrolides have a broad spectrum of activity, but are primarily used for pertussis and mycoplasma coverage. In terms of adverse effects, erythromycin and clitorithromycin have multiple drug-drug interactions that need to be considered, and all macrolides can prolong the QT interval. Again, need to be very careful in patients who may be receiving other drugs that can prolong the QT interval or in other patients who are at risk for QT prolongation. Rifamicin are often used as adjunctive treatment for MRSA or penicillin-resistant streptococcal infections. They're almost always used in combination with other antibiotics because of the high propensity for organisms to develop resistance. They can also be used for gut decontamination in the setting of hepatic encephalopathy. Rifamicin have multiple drug interactions that need to be taken into account. They can cause hepatotoxicity, bone marrow suppression, and interstitial nephritis. They will also cause body fluids to turn red, which can be concerning to patients if they're not aware of this. Fluoroquinolones have a broad range of antimicrobial activity. They're often not the first-line agents in pediatrics, though they can be helpful as an oral antipsydumonal agent or for complicated UTIs or pyelonephritis. They can also be helpful as empiric antipsydumonous coverage in patients with kidney disease for whom you might want to limit aminoglycoside exposure. Importantly, as a side effect, fluoroquinolones can cause QT prolongation. Regarding sulfa antibiotics, trimethoprim sulfamethoxazole has activity against gram-positive cocci and some enteric gram-negative organisms, though it's not recommended for strep pyogenes infections due to high resistance patterns. It can be used to treat MRSA and is the drug of choice for pneumocystis and stenotropomonas infections. Adverse effects include myelosuppression and hypersensitivity reactions. Rounding out our discussion of antibiotics, it's important to consider pharmacodynamics when thinking about dosing strategies and therapeutic drug monitoring. Some antibiotics work in a concentration-dependent manner, whereby the concentration of the antibiotic and the peak serum concentrations relative to the MIC are important factors that determine effective killing. Other antibiotics work in a time-dependent manner, whereby the amount of time spent at a certain concentration is important. These considerations are important not just for dosing, but for therapeutic drug monitoring, and particularly in a setting of a critically ill patient that may have drug concentrations that are harder to predict based on dosing strategies because of differences in fluid shifts and differences in organ function over time. I'll spend the next couple of slides talking about antimicrobial stewardship. The goal of these programs is to reduce the misuse of antibiotics, tying together healthcare value with patient safety. These programs are vitally important to stem the tide of emerging drug-resistant organisms. The Infectious Disease Society of America in 2007 developed guidelines for antimicrobial stewardship programs. These guidelines were revised in 2016. Important features include a multidisciplinary team, as well as some suggestions for stewardship programming, including suggested or scripted forms. Although novel forms or forms that are derived and implemented within individual institutions have also had good results. As with most QI initiatives, the success of a formal antimicrobial stewardship program depends on buy-in from key stakeholders. This includes hospital administration and the individuals who are prescribing antibiotics. It also involves having resources that allow one to track outcomes and disseminate the information. Some success stories include reduction in all antibiotic use, reduction in targeted antibiotics, decreased cost, and importantly, shifts in resistance patterns. Across multiple studies, no single ASP method has been proven superior. But in general, having an ASP seems to be beneficial compared to not having one, and so the important piece is to get started. For more information, this is a paper that describes a multi-center collaborative of 36 US children's hospitals that have formed to share best practices, including sharing forms and sharing data reporting tools in support of antimicrobial stewardship programs. We'll round out the presentation with a discussion of four cases. These four different cases will highlight different types of life-threatening infection in critically ill children. We start with case 1. As you'll recall, case 1 was a seven-month-old unimmunized female admitted with fever, emesis, and irritability. Significant physical exam findings included a tense, bulging anterior fontanelle, a stiff neck, tachypnea, tachycardia, and irritability. Lab findings were as we reviewed previously, although now you have CSF results available to you. That CSF is, of course, consistent with meningitis with elevated white blood cell count, a low glucose, high protein, and a gram stain with many white blood cells and many gram-negative bacilli. So the question here again is, what is the most likely organism and what antibiotics will you choose? So ordinarily for this age group, we would consider strep pneumo and Neisseria meningitidis as our top two potentially causative organisms for meningitis. But recall this was an unimmunized infant, and so Haemophilus influenzae is also on the list, and most likely those gram-negative rods in the CSF are Haemophilus. In terms of empiric antibiotic therapy, this table lists recommended empiric antibiotics when there is a concern for meningitis based on the age immunization status of the patient. For infants under one month of age, one would consider ampicillin to cover Listeria plus cefetaxime or gentamicin. For all other age groups, empiric therapy would include a third-generation cephalosporin plus vancomycin, as is the case for our infant. Our second case is a 14-year-old admitted with a two-day history of fever, vomiting, diarrhea, body aches, progressively worsening headache, confusion, and an erythematous rash. It was noted that two weeks prior to admission, he fell off his skateboard and sustained a deep laceration to his right knee. On physical exam, he is sleepy and confused, but complaining of body aches and headache. He's tachycardic, hypotensive, tachypneic, has poor cap refill, and has a right knee that is swollen with a healing laceration with surrounding induration, overlying erythema, purulent discharge, and tender to palpation. He has mental status changes and a diffuse erythematous confluent macular rash. Laboratory findings are notable for leukocytosis, thrombocytopenia, elevated ESR-CRP, mild hyponatremia, and elevated buencreatinine, as well as an elevated ASD-ALT. Gram stain from the laceration shows gram-positive cocci, as well as many white blood cells. The x-ray of the right knee is concerning for potential retained foreign body, periosteal reaction, and osteolytic lesion of the distal femur. So the question is, what is the most appropriate empiric antibiotic regimen for this patient? This patient is exhibiting signs and symptoms of toxic shock syndrome. Toxic shock syndrome is caused by staphylococcal or streptococcal bacterial toxins that act as super antigens. These super antigens cause nonspecific T-cell activation that results in an overwhelming inflammatory response. Diagnostic criteria for toxic shock syndrome differ based on whether the causative organism is staph or strep. The diagnostic criteria for staphylococcal toxic shock syndrome are listed here. Of note, they include fever, hypotension, a diffuse macular rash with subsequent desquamation, and involvement of three or more organ systems. For streptococcal toxic shock syndrome, diagnostic criteria include hypotension and two or more of the following organ systems being involved. They also involve isolation of group A strep from either a sterile site for a definitive case or a non-sterile site for a probable case. Risk factors for toxic shock syndrome include colonization with toxin-producing staph aureus, absence of protective antitoxin antibodies, and primary staph aureus or strep pyogenes infection. These infections can result from really anything that causes skin or mucous membrane disruption. This can include trauma, burns, insect bites, needle sticks, or certain viral infections. They can also involve surgical or non-surgical foreign bodies. Management of toxic shock syndrome involves early appropriate antibiotic therapy, which is of paramount importance. Empiric antibiotics should include gram-positive coverage, including MRSA, until culture results are available. The addition of a protein synthesis inhibitor, such as clindamycin, can be useful in order to decrease protein or toxin production. Source control is also vitally important, particularly in the setting of deep-seated infection or foreign body. Hemodynamic support and management of shock and multiple organ dysfunction is often necessary. IVIG can be added to provide neutralizing antibodies to counter the effects of bacterial toxins, and IVIG use has been associated with improved mortality in the setting of toxic shock syndrome. Our third case is of a 14-year-old previously healthy child who presents in February with a five-day history of fever, body aches, shaking chills, worsening shortness of breath, and cough. These symptoms rapidly progress to respiratory failure, ultimately requiring ECMO support. It's noted that his household contacts have been ill with fever and URI symptoms, and though he is up to date with his immunizations, he did not receive an influenza vaccine and community rates of influenza infection are high. So what empiric antibiotics would you choose for this child? To answer this question, we need to consider bacterial co-infection in the setting of influenza. These are data from the PICFLU multi-center observational study of 170 critically ill children with influenza across 34 centers. Within the cohort, 53% had evidence of bacterial co-infection. Of those, about a third were MRSA, about a third were MSSA, and about a third were other bacteria. Importantly, mortality rates among those with influenza plus MRSA were much higher than those with either non-MRSA bacteria or no bacteria identified. For critically ill children with influenza and suspected bacterial infection, empiric antibiotic therapy should include MRSA coverage. For those with confirmed MRSA and influenza, there may be a role for combination therapy, though studies are small. These are data from one of those studies that suggested an association between vancomycin monotherapy and non-survival in the setting of influenza and MRSA. Though these data are not definitive, it is reasonable to monitor closely for treatment failure and involve your infectious disease colleagues early given the high-risk nature of these patients. Our fourth case is a 15-year-old female with a two-day history of fever, worsening sore throat, left-sided neck swelling and stiffness, extreme fatigue and headache. Over the last 24 hours, she has had several episodes of emesis without URI symptoms and is starting to complain of feeling lightheaded and having shortness of breath. Physical exam findings are significant for shortness of breath, fever, tachycardia, tachypnea, hypotension, low oxygen saturation on remer, a relatively normal pharyngeal exam, as well as left-sided neck swelling that is tender to palpation. Lungs have scattered ronchi and she has tachypneic and she has tachycardic, as mentioned, with a delayed cap refill. Laboratory findings reveal leukocytosis, mild thrombocytopenia, and an elevated CRP, as well as a positive blood culture. On imaging, there are several large ill-defined nodular opacities on chest radiograph NCT and a thrombus in the internal jugular vein noted on neck CT. This is a picture of the chest X-ray here where you can see multifocal airspace opacities. And the chest and neck CT findings, in the neck CT, the arrow is pointing to an internal jugular vein thrombosis. The chest CT reveals multiple septic emboli. And so the question is, what is the diagnosis and what is the most likely organism? This is a case of Lemire syndrome. Lemire syndrome was described in the 30s by Andre Lemire. It's characterized by acute oropharyngeal infection, bacteremia, septic thrombophlebitis of the internal jugular and or facial veins, and distant septic emboli, most commonly to the lungs, though other organ systems can also be involved. The most common bacteria is Fusobacterium, which is a gram-negative anaerobe. There are two peaks of incidence in terms of age, one in adolescents or young adults and another in the elderly. In terms of adolescents or young adults, the upper respiratory tract and head and neck area are the primary sites of infection. Lemire syndrome has a fairly high mortality rate despite treatment ranging from 4% to 25%. Common clinical findings include high fevers, chills, oropharyngeal pain, ipsilateral neck swelling and tenderness, parallel to the sternocleidomastoid muscle, and oftentimes abdominal pain as well. The diagnosis is based on high degree of clinical suspicion as well as the presence of the aforementioned clinical signs and a positive blood culture. CT or Doppler ultrasound may be necessary to evaluate for internal jugular vein thrombosis. Chest X-ray and high resolution CT may also be used to evaluate for septic emboli in the lungs. It's also important to consider that embolic abscesses to the brain, kidneys, bone, joints, and other locations may also be seen and should be evaluated for. In terms of laboratory tests, a CBC will show elevated white count with a left shift, ESR and CRP will be elevated, and coagulation panels may show mild DIC. Typically, blood cultures are positive. Prolonged treatment is often needed because of the septic embolization with a mean duration of therapy of about six weeks. While penicillins may be effective, there are some concerns about the emergence of beta-lactamase producing organisms, and so the recommended therapy is clindamycin, metronidazole, or penicillin with a beta-lactamase inhibitor. These patients often require a prolonged anticoagulant course as well for severe septic thrombophlebitis with a mean duration of up to three months. Surgical intervention for excision of internal jugular vein may also be required if antibiotics and or anticoagulants are ineffective. In summary, life-threatening infection is a common cause for PICU admission. A thoughtful approach to antibiotic therapy is needed in order to balance the needs for rapid administration of appropriate antibiotic coverage with avoidance of unnecessary antibiotics and worsening antibiotic resistance. Antibiotic selection depends on the likely causative organisms and sites of infection, and these can be influenced by the age and immunization status of your patient, the geography and your local patterns of antibiotic resistance, and the presence of complex chronic conditions, including immune compromise and frequency of contact with the healthcare system.

life-threatening infections

pediatric intensive care unit

antibiotic therapy

inappropriate prescriptions

causative organism

antimicrobial stewardship programs

antibiotic resistance