Peter K. Linden

Guidelines for evaluation of new fever in critically ill adult patients: 2008 update from the American College of Critical Care Medicine and the Infectious Diseases Society of America

Objective:To update the practice parameters for the evaluation of adult patients who develop a new fever in the intensive care unit, for the purpose of guiding clinical practice. Participants:A task force of 11 experts in the disciplines related to critical care medicine and infectious diseases was convened from the membership of the Society of Critical Care Medicine and the Infectious Diseases Society of America. Specialties represented included critical care medicine, surgery, internal medicine, infectious diseases, neurology, and laboratory medicine/microbiology. Evidence:The task force members provided personal experience and determined the published literature (MEDLINE articles, textbooks, etc.) from which consensus was obtained. Published literature was reviewed and classified into one of four categories, according to study design and scientific value. Consensus Process:The task force met twice in person, several times by teleconference, and held multiple e-mail discussions during a 2-yr period to identify the pertinent literature and arrive at consensus recommendations. Consideration was given to the relationship between the weight of scientific evidence and the strength of the recommendation. Draft documents were composed and debated by the task force until consensus was reached by nominal group process. Conclusions:The panel concluded that, because fever can have many infectious and noninfectious etiologies, a new fever in a patient in the intensive care unit should trigger a careful clinical assessment rather than automatic orders for laboratory and radiologic tests. A cost-conscious approach to obtaining cultures and imaging studies should be undertaken if indicated after a clinical evaluation. The goal of such an approach is to determine, in a directed manner, whether infection is present so that additional testing can be avoided and therapeutic decisions can be made.

Opportunistic Mycelial Fungal Infections in Organ Transplant Recipients: Emerging Importance of Non-Aspergillus Mycelial Fungi

To determine the spectrum and impact of mycelial fungal infections, particularly those due to non-Aspergillus molds, 53 liver and heart transplant recipients with invasive mycelial infections were prospectively identified in a multicenter study. Invasive mycelial infections were due to Aspergillus species in 69.8% of patients, to non-Aspergillus hyalohyphomycetes in 9.4%, to phaeohyphomycetes in 9.4%, to zygomycetes in 5.7%, and to other causes in 5.7%. Infections due to mycelial fungi other than Aspergillus species were significantly more likely to be associated with disseminated (P=.005) and central nervous system (P=.07) infection than were those due to Aspergillus species. Overall mortality at 90 days was 54.7%. The associated mortality rate was 100% for zygomycosis, 80% for non-Aspergillus hyalohyphomycosis, 54% for aspergillosis, and 20% for phaeohyphomycosis. Thus, non-Aspergillus molds have emerged as significant pathogens in organ transplant recipients. These molds are more likely to be associated with disseminated infections and to be associated with poorer outcomes than is aspergillosis.

Determinants of Vancomycin Resistance and Mortality Rates in Enterococcal Bacteremia: A Prospective Multicenter Study

Enterococcus species have become increasingly prominent as etiologic agents of nosocomial bacteremia (19). Enterococcal bacteremia has a mortality rate of 42% to 73% (10, 11) and is common among debilitated patients and those with severe underlying illnesses (5, 6, 1217). Enterococci have low-level resistance to penicillins, aminoglycosides, and clindamycin and are intrinsically resistant to cephalosporins. Enterococci may acquire resistance to additional antibiotics, including -lactams, aminoglycosides, and glycopeptides (18). Resistance to multiple antibiotics, in particular vancomycin coupled with high-level ampicillin and aminoglycoside, has been reported with increasing frequency (19). At present, more than 20% of enterococci isolated from intensive care units exhibit vancomycin resistance. The addition of vancomycin resistance to high-level ampicillin and aminoglycoside resistance limits available therapeutic options (20). To investigate the clinical implications of antibiotic resistance in enterococci, we instituted a prospective, multicenter observational study of outcome in patients with enterococcal bacteremia. We sought to determine 1) factors associated with infection with vancomycin-resistant enterococci [VRE], 2) factors predictive of death in patients with enterococcal bacteremia, 3) the effect of vancomycin resistance on mortality rates, and 4) the effect of antibiotic therapy on outcome. Methods All patients with enterococcal bacteremia were hospitalized at the University of Pittsburgh Medical Center and the Veterans Affairs (VA) Medical Center (Pittsburgh, Pennsylvania), Detroit Medical Center and John D. Dingell VA Medical Center (Detroit, Michigan), Rush-Presbyterian-St. Lukes Medical Center (Chicago, Illinois), New England Medical Center (Boston, Massachusetts), and William Beaumont Hospital (Royal Oak, Michigan). Clinical data were obtained from review of medical records. The institutional review boards of four participating institutions approved the study. At the fifth institution, the study was considered exempt from review because it involved confidential use of existing records and bacterial isolates. Microbiology Blood for culture was obtained by venipuncture or through central venous catheters. Enterococcal species were determined by using either VITEK (bioMerieux Vitek, Inc., Hazelwood, Missouri) or MicroScan (MicroScan, Inc., West Sacramento, California) systems according to the manufacturers recommendations. Identification of species other than E. faecalis and E. faecium was confirmed as reported elsewhere (21). One of the authors standardized antimicrobial susceptibilities by using Etest strips (AB BIODISK North America, Inc., Piscataway, New Jersey). Enterococcus faecium isolates showing resistance or intermediate susceptibility to quinupristindalfopristin were tested by broth microdilution and disk diffusion to confirm reduced susceptibility. If the isolate was unavailable, antimicrobial susceptibilities reported by the submitting microbiology laboratories were used. Minimum inhibitory concentration (MIC) breakpoints from the Ninth National Committee for Clinical Laboratory Standards (NCCLS) were used (22). Because imipenem MIC values for enterococci are not defined, NCCLS breakpoints for Enterobacteriaceae were used (22). Quality control was monitored by using E. faecalis American Type Culture Collection (ATCC) 29212. Nine enterococcal isolates displayed intermediate susceptibility to vancomycin (MIC, 8 to 16 g/mL) but were considered resistant for the purposes of analysis. Vancomycin resistance genotypes of selected clinical isolates were determined by using polymerase chain reaction (PCR) amplification with primers specific for intragenic sequences of the vanA and vanB genes (23, 24). Control strains included vancomycin-susceptible E. faecalis ATCC 29212, E. faecium BM4147 (vanA) (25), and E. faecalis V583 (vanB) (26). For determination of aminoglycoside resistance genes, genomic DNA for PCR amplification was prepared with the InstaGene Matrix kit (BioRad Laboratories, Hercules, California) and PCR performed as reported elsewhere (27). Aminoglycoside resistance genes detected included aac(6)-Ie-aph(2)-Ia (28), aph(2)-Ic (29), aph(2)-Id (30), and aph(2)-Ib (31). Definitions Clinically significant bacteremia was defined as isolation of enterococci in two or more separately obtained blood cultures or from a single blood culture and from a concomitant site of infection in a clinical scenario compatible with bacteremic infection (6). Endocarditis was defined by using the Duke criteria (32). Polymicrobial bacteremia was defined as isolation from blood culture of one or more additional species of bacteria concomitantly with enterococci (same blood culture or another blood culture within 24 hours of the initial blood culture yielding enterococci). A single concomitant isolation of another bacterial species was sufficient, except for isolation of coagulase-negative staphylococci, diphtheroids, -hemolytic streptococci, and Bacillus species that required isolation from two blood cultures. Length of hospitalization was defined as the time in days from hospital admission to development of clinically significant enterococcal bacteremia. Enterococcal bacteremia occurring 60 days or more from a previous episode in patients already enrolled was counted as a separate episode. The end point was survival at 14 days from the first positive blood culture. Patients discharged from the hospital before 14 days were considered survivors. Medical records were reviewed at entry, at 2 weeks, at 4 weeks, and at 6 weeks (or at time of discharge or death if earlier than 6 weeks). We collected information on patient demographic characteristics, underlying disease, Acute Physiology and Chronic Health Evaluation (APACHE) II scores at bacteremia onset, antibiotic use, use of glucocorticosteroids and other immunosuppressive drugs, and receipt of invasive devices and procedures in the 2 weeks before bacteremia onset. Antibiotic therapy during the 6 weeks after the onset of bacteremia was recorded. Beyond 6 weeks, patients were followed for evidence of relapse of bacteremia and for survival to discharge or death. Immunosuppressive drugs included cyclosporine, cyclophosphamide, azathioprine, tacrolimus, methotrexate, and cytotoxic chemotherapy. Appropriate antibiotic therapy was defined as treatment with at least one antibiotic that had in vitro activity (as defined by Etest) against the enterococcal isolate, initiated within 48 hours of the initial positive enterococcal blood culture and continuing for at least 72 hours. Antibiotics considered potentially active included penicillin, ampicillin, ureidopenicillin, vancomycin, quinupristindalfopristin, chloramphenicol, doxycycline, and rifampin. Statistical Analysis For categorical variables, proportions were compared by using the Fisher exact test. Continuous variables were analyzed with the MannWhitney rank-sum test. Multivariate analysis was done by using logistic regression. Variables with a two-tailed P value of 0.05 were included in stepwise logistic regression models for vancomycin resistance and 14-day mortality. The initial bacteremic episode for each patient (n = 398) was used for the evaluation of risk factors for bacteremia caused by VRE. Patients who were alive 14 days after the onset of enterococcal bacteremia (n = 321) were evaluated for factors associated with microbiological failure. Statistical analysis of the data was performed by using the Prophet system (MarketMiner, Inc., Charlottesville, Virginia) and Epistat (Epistat Services, Richardson, Texas). Results Enterococcal Bacteremia We studied hospitalized patients 16 years of age or older with clinically significant hospital- or community-acquired enterococcal bacteremia. From February 1995 through March 1997, 391 consecutive patients from five participating institutions were entered into the study. An additional 9 patients from Pittsburgh were entered into the study over a 6-month period (October 1998 through March 1999) to increase the total number of patients to 400. These patients were consecutive and unselected. Two patients younger than 16 years of age were excluded, leaving 398 patients for evaluation. Eighty-nine patients were from the University of Pittsburgh Medical Center and the VA Medical Center, 97 were from Rush-Presbyterian-St. Lukes Medical Center, 95 were from the Detroit Medical Center and John D. Dingell VA Medical Center, 61 were from New England Medical Center, and 56 were from the William Beaumont Hospital. Blood cultures yielded 398 enterococcal isolates. Of these, 60% (239 of 398) were E. faecalis and 37% (148 of 398) were E. faecium. Three percent (10 of 398) of the isolates belonged to the less common enterococcal species, which include E. avium, E. casseliflavus, E. durans, E. gallinarum, and E. raffinosus. The species of one isolate could not be identified. Seventeen recurrences were seen at 60 or more days after the initial bacteremia. Of these, 14 were caused by the same enterococcal species as the initial episode. In 12 of these 14 episodes, the pair of isolates had the same susceptibilities to vancomycin. Overall, 35% of the 398 enterococcal isolates were resistant to vancomycin (MIC 32 g/mL), 63% were susceptible to vancomycin (MIC 4 g/mL), and 2% displayed intermediate susceptibility (MIC, 8 to 16 g/mL). Table 1 shows the susceptibility patterns of the two major Enterococcus species. Eight percent of the E. faecalis isolates were resistant to vancomycin, 91% were susceptible, and 1% displayed intermediate susceptibility. In contrast, 80% of the E. faecium isolates were resistant to vancomycin, 18% were susceptible, and 2% displayed intermediate susceptibility. Of the 11 isolates of other species, none were resistant to vancomycin, 73% (8 of 11) were susceptible to vancomycin, and 27% (3 of 11) displayed intermediate susceptibility to vancomycin. Table 1. Pr

Posttransplant Lymphoproliferative Disease in Primary Epstein-Barr Virus Infection after Liver Transplantation: The Role of Cytomegalovirus Disease

Epstein-Barr virus (EBV) plays a major role in the pathogenesis of posttransplant lymphoproliferative disease (PTLD). Patients who undergo primary EBV infection after transplantation are at greater risk of developing PTLD. In this retrospective study, the incidence of EBV infection and associated PTLD in 40 consecutive adult recipients who were seronegative for EBV at the time of liver transplantation were investigated, and risk factors for PTLD were analyzed. Of 37 patients with available timely posttransplant serum samples, 35 (95%) developed primary EBV infection. Of the 40 patients, 13 (33%) developed PTLD a median of 126 days (range, 48-776) after liver transplantation. The factor significantly associated with the development of PTLD was cytomegalovirus disease (relative risk, 7.3; 95% confidence interval, 2.36-22.6; P = .0006). Cytomegalovirus disease is a predictor for the development of PTLD in primary EBV infection after liver transplantation, and it may be a target for prophylactic intervention.

Use of Parenteral Colistin for the Treatment of Serious Infection Due to Antimicrobial-Resistant Pseudomonas aeruginosa

Serious infection due to strains of Pseudomonas aeruginosa that exhibit resistance to all common antipseudomonal antimicrobials increasingly is a serious problem. Colistin was used as salvage therapy for 23 critically ill patients with multidrug-resistant P. aeruginosa infection. Twenty-two patients who had septic shock (n=14) and/or renal failure (n=21) received mechanical ventilatory support at baseline. The most common types of infection were pneumonia (n=18) and intra-abdominal infection (n=5). Colistin was administered for a median of 17 days (range, 7-36 days). Seven patients died during therapy, at a median of 17 days (range, 4-26 days) after initiation of treatment. A favorable clinical response was observed in 14 patients (61%); only 3 patients experienced relapse. Bacteremia was the only significant factor associated with treatment failure (P=.02). One patient manifested diffuse weakness that resolved after temporary cessation of colistin therapy. Colistin provides an important salvage therapeutic option for patients with otherwise untreatable serious P. aeruginosa infection.

Zygomycosis in Solid Organ Transplant Recipients in a Tertiary Transplant Center and Review of the Literature

Zygomycetes are ubiquitous fungi that can cause invasive disease associated with high mortality. We report 10 solid organ transplant recipients with zygomycosis (incidence 2 per 1000) and reviewed 106 cases in the English‐language literature. These included renal (n = 73), heart (n = 16), lung (n = 4), heart/lung (n = 2), liver (n = 19) and kidney/pancreas (n = 2) transplant recipients. All patients were receiving immunosuppression and the vast majority steroids. The clinical presentation included rhino‐sino‐orbital (n = 20), rhinocerebral (n = 16), pulmonary (n = 28), gastrointestinal (n = 13), cutaneous (n = 18), renal (n = 6) and disseminated disease (n = 15). Most frequently isolated genera were Rhizopus (73%) followed by Mucor (13%). The overall mortality was 49%. While rhino‐sino‐orbital disease had the best prognosis, rhinocerebral disease had high mortality (93%) comparable to disseminated disease. A favorable outcome was associated with limited, surgically accessible disease and early surgical intervention along with amphotericin B administration.

Practice Parameters for Evaluating New Fever in Critically Ill Adult Patients. Task Force of the American College of Critical Care Medicine of the Society of Critical Care Medicine in Collaboration with the Infectious Disease Society of America.

Abstract Objective: To develop practice parameters for the evaluation of adult patients who develop a new fever in the intensive care unit (ICU) for the purpose of guiding clinical practice. Participants: A task force of 13 experts in disciplines related to critical care medicine, infectious diseases, and surgery was convened from the membership of the Society of Critical Care Medicine, and the Infectious Disease Society of America. Evidence: The task force members provided the personal experience and determined the published literature (MEDLINE articles, textbooks, etc.) from which consensus would be sought. Published literature was reviewed and classified into one of four categories, according to study design and scientific value. Consensus Process: The task force met several times in person and twice monthly by teleconference over a 1‐yr period of time to identify the pertinent literature and arrive at consensus recommendations. Consideration was given to the relationship between the weight of scientific evidence and the experts’ opinions. Draft documents were composed and debated by the task force until consensus was reached by nominal group process. Conclusions: The panel concluded that, because fever can have many infectious and noninfectious etiologies, a new fever in a patient in the ICU should trigger a careful clinical assessment rather than automatic orders for laboratory and radiologic tests. A cost‐conscious approach to obtaining cultures and imaging studies should be undertaken if it is indicated after a clinical evaluation. The goal of such an approach is to determine, in a directed manner, whether or not infection is present, so additional testing can be avoided and therapeutic options can be made. (Crit Care Med 1998; 26:392‐408) In some intensive care units (ICUs), the measurement of a newly elevated temperature triggers an automatic order set which includes many tests that are time consuming, costly, and disruptive (Table 1). Moreover, the patient may experience discomfort, be exposed to unneeded radiation, or experience considerable blood loss due to this testing, which is often repeated several times within 24 hrs, and daily thereafter. In an era when utilization of hospital and patient resources is under intensive scrutiny, it is appropriate to assess how such fevers should be evaluated in a prudent and cost‐effective manner. Table 1. Typical costs associated with fever evaluation The American College of Critical Care Medicine of the Society of Critical Care Medicine and the Infectious Disease Society of America established a Task Force to provide practice parameters for the evaluation of a new fever in patients in an ICU with the goal of promoting the rational consumption of resources and promoting an efficient evaluation. These practice parameters presume that any unexplained temperature elevation merits a clinical assessment by a healthcare professional that includes a review of the patients history and a focused physical examination before any laboratory tests or imaging procedures are ordered. These practice parameters specifically address how to evaluate a new fever in an adult patient already in the ICU who has previously been afebrile and in whom the source of fever is not initially obvious. If the initial evaluation of history and physical examination reveals a consolidated lung, a purulent wound, or a phlebitic leg, then diagnosis and therapy of that infectious process should commence: such management is addressed by other practice parameters aimed specifically at pneumonia, catheter‐related infections, etc. Specific questions addressed in these practice parameters relate to the search for the underlying cause of fever and include: a) What temperature should elicit an evaluation? b) When are blood cultures warranted? c) When should intravascular catheters be cultured or removed? d) When are cultures of respiratory secretions, urine, stool, or cerebral spinal fluid warranted? e) When are radiographic studies warranted? These practice parameters do not address children, since children have different issues that merit discussion in a separate document. In addition, these practice parameters do not address an approach to persistent fever after the initial evaluation, or to localized infection once the anatomic source of fever has been identified. These issues are addressed in other monographs or practice parameters. The current document also does not address the desirability or selection of empiric vs. specific therapy since the need for therapy is so dependent on clinical evaluation and the underlying disease. It did not appear to this task force that useful therapeutic guidelines could easily be provided which took into account the acuity of illness, the underlying disease process, concurrent drugs (i.e., immunosuppressive agents, and antimicrobials), ability to tolerate toxicities, and geographic antibiotic susceptibility differences. Each ICU must establish its own policies for evaluating fever that take into account the type of ICU involved (e.g., medical ICU, surgical ICU, burn ICU, etc.), the specific patient population (e.g., immunosuppressed vs. immunocompetent, elderly vs. younger adults), recent epidemics (e.g., out‐breaks of Clostridium difficile diarrhea or vancomycin‐resistant Enterococcus), or endemic pathogens (e.g., methicillin‐resistant Staphylococcus aureus). It is hoped that these practice parameters will assist intensivists and consultants as a starting point for developing an effective and cost conscious approach appropriate for their patient populations. The specific recommendations are rated by the strength of evidence, using the published criteria of the Society of Critical Care Medicine (Table 2). Table 2. Society of Critical Care Medicines rating system for strength of recommendation and quality of evidence supporting the references

Treatment Options for Vancomycin-Resistant Enterococcal Infections

Serious infection with vancomycin-resistant enterococci (VRE) usually occurs in patients with significantly compromised host defences and serious comorbidities, and this magnifies the importance of effective antimicrobial treatment. Assessments of antibacterial efficacy against VRE have been hampered by the lack of a comparator treatment arm(s), complex treatment requirements including surgery, and advanced illness-severity associated with a high crude mortality.Treatment options include available agents which don’t have a specific VRE approval (chloramphenicol, doxycycline, high-dose ampicillin or ampicillin/sulbactam), and nitrofurantoin (for lower urinary tract infection). The role of antimicrobial combinations that have shown in vitro or animal-model in vivo efficacy has yet to be established. Two novel antimicrobial agents (quinupristin/dalfopristin and linezolid) have emerged as approved therapeutic options for vancomycin-resistant Enterococcus faecium on the basis of in vitro susceptibility and clinical efficacy from multicentre, pharmaceutical company-sponsored clinical trials.Quinupristin/dalfopristin is a streptogramin, which impairs bacterial protein synthesis at both early peptide chain elongation and late peptide chain extrusion steps. It has bacteriostatic activity against vancomycin-resistant E. faecium [minimum concentration to inhibit growth of 90% of isolates (MIC90) = 2 μg/ml] but is not active against Enterococcus faecalis (MIC90= 16μg/ml). In a noncomparative, nonblind, emergency-use programme in patients who were infected with Gram-positive isolates resistant or refractory to conventional therapy or who were intolerant of conventional therapy, quinupristin/dalfopristin was administered at 7.5 mg/kg every 8 hours. The clinical response rate in the bacteriologically evaluable subset was 70.5%, and a 65.8% overall response (favourable clinical and bacteriological outcome) was observed. Resistance to quinupristin/dalfopristin on therapy was observed in 6/338 (1.8%) of VRE strains. Myalgia/arthralgia was the most frequent treatment-limiting adverse effect. In vitro studies which combine quinupristin/dalfopristin with ampicillin or doxycyline have shown enhanced killing effects against VRE; however, the clinical use of combined therapy remains unestablished.Linezolid, an oxazolidinone compound that acts by inhibiting the bacterial pre-translational initiation complex formation, has bacteriostatic activity against both vancomycin resistant E. faecium (MIC90 = 2 to 4 μg/ml) and E. faecalis (MIC90 = 2 to 4 μg/ml). This agent was studied in a similar emergency use protocol for multi-resistant Gram-positive infections. 55 of 133 evaluable patients were infected with VRE. Cure rates for the most common sites were complicated skin and soft tissue 87.5% (7/8), primary bacteraemia 90.9% (10/11), peritonitis 91.7% (11/12), other abdominal/pelvic infections 91.7% (11/12), and catheter-related bacteraemia 100% (9/9). There was an all-site response rate of 92.6% (50/54).In a separate blinded, randomised, multicentre trial for VRE infection at a variety of sites, intravenous low dose linezolid (200mg every 12 hours) was compared to high dose therapy (600mg every 12 hours) with optional conversion to oral administration. A positive dose response (although statistically nonsignificant) was seen with a 67% (39/58) and 52% (24/46) cure rate in the high- and low-dose groups, respectively.Adverse effects of linezolid therapy have been predominantly gastrointestinal (nausea, vomiting, diarrhoea), headache and taste alteration. Reports of thrombocytopenia appear to be limited to patients receiving somewhat longer courses of treatment (>14 to 21 days). Linezolid resistance (MIC ≥8 μg/ml) has been reported in a small number of E. faecium strains which appears to be secondary to a base-pair mutation in the genome encoding for the bacterial 23S ribosome binding site. At present a comparative study between the two approved agents for VRE (quinuprisin/dalfopristin and linezolid) has not been performed.Several investigational agents are currently in phase II or III trials for VRE infection. This category includes daptomycin (an acidic lipopeptide), oritavancin (LY-333328; a glycopeptide), and tigilcycline (GAR-936; a novel analogue of minocycline). Finally, strategies to suppress or eradicate the VRE intestinal reservoir have been reported for the combination of oral doxycyline plus bacitracin and oral ramoplanin (a novel glycolipodepsipeptide). If successful, a likely application of such an approach is the reduction of VRE infection during high risk periods in high risk patient groups such as the post-chemotherapy neutropenic nadir or early post-solid abdominal organ transplantation.

Association between the Presence of Enterococcal Virulence Factors Gelatinase, Hemolysin, and Enterococcal Surface Protein and Mortality among Patients with Bacteremia Due to Enterococcus faecalis

The potential virulence factors of enterococci include production of enterococcal surface protein (Esp), gelatinase, and hemolysin. Gelatinase- and hemolysin-producing strains of Enterococcus faecalis have been shown to be virulent in animal models of enterococcal infections. Esp production has been shown to enhance the persistence of E. faecalis in the urinary bladder. We determined the presence of the esp gene and production of gelatinase and hemolysin in 219 E. faecalis isolates from a larger prospective study of 398 patients with enterococcal bacteremia. Thirty-two percent of isolates carried the esp gene, 64% produced gelatinase, and 11% produced hemolysin. There was no significant association between 14-day mortality and any of the markers studied, singly or in combination.

A PROSPECTIVE RANDOMIZED TRIAL COMPARING SEQUENTIAL GANCICLOVIR-HIGH DOSE ACYCLOVIR TO HIGH DOSE ACYCLOVIR FOR PREVENTION OF CYTOMEGALOVIRUS DISEASE IN ADULT LIVER TRANSPLANT RECIPIENTS

Cytomegalovirus disease is an important cause of morbidity following liver transplantation. To date there has not been an effective prophylaxis for CMV disease after liver transplantation. One hundred forty-three patients were randomized to receive either high dose oral acyclovir (800 mg 4 times a day) alone for 3 months after transplantation (acyclovir group) or intravenous ganciclovir (5 mg/kg twice a day) for 14 days followed by high dose oral acyclovir to complete a 3-month regimen (ganciclovir group). Of 139 patients available for evaluation, 43 of 71 (61%) patients from the acyclovir group developed CMV infection compared with 16 of 68 (24%) from the ganciclovir group (relative risk, 3.69; 95% confidence interval, 2.07–6.56; P<0.00001). Of those randomized, CMV disease was seen in 20 (28%) of the acyclovir group compared with 6 (9%) of the ganciclovir group (relative risk, 5.11; 95% confidence interval, 2.05–12.75; P=0.0001). The median time to onset of CMV infection was 45 days in the acyclovir group compared with 78 days in the ganciclovir group (P=0.004). The median time to onset of CMV disease was 40 days in the acyclovir group compared with 78 days in the ganciclovir patients (P=0.02). With respect to primary CMV infection, there was no difference in the rates in the 2 groups, but tissue invasive disease and recurrent CMV disease were less frequent in the ganciclovir group. It is concluded that a course of 2 weeks of ganciclovir immediately after transplantation followed by high dose oral acyclovir for 10 weeks is superior to a 12-week course of high dose oral acyclovir alone for prevention of both CMV infection and CMV disease after liver transplantation. However, the lack of significant effect in sero-negative recipients who received grafts from sero-positive donors suggests that other strategies are needed to prevent CMV infection in this high risk population.