The costs of hospital-acquired (nosocomial) and other health care–associated infections are great. It is estimated that these infections affect >2 million patients, cost $4.5 billion, and contribute to 88,000 deaths in U.S. hospitals annually. Efforts to lower infection risks have been challenged by the growing numbers of immunocompromised patients; antibiotic-resistant bacteria, fungal, and viral superinfections; and invasive devices and procedures. Nevertheless, evidence-based guidelines for prevention and control are available (Table 13-1); according to some estimates, consistent application of these guidelines may reduce the risk of health care–associated infection by more than one-third, and the growing viewpoint of consumer advocates is that almost all such infections are preventable. This chapter reviews health care–acquired and device-related infections and the basic surveillance, prevention, control, and treatment activities that have been developed to deal with these problems.


The standards of the Joint Commission require all accredited hospitals to have an active program for surveillance, prevention, and control of nosocomial infections. Education of physicians in infection control and health care epidemiology is required in infectious disease fellowship programs and is available by online courses. Diagnosis-related reimbursement has led hospital administrators to place increased emphasis on infection control. Federal concerns over “patient safety” have led to legislation that would limit reimbursement for hospital costs resulting from at least two (yet-to-be-determined) nosocomial infections. The patient safety movement has prompted major national efforts to improve, measure, and publicly report on processes of patient care (e.g., timely administration and appropriateness of perioperative antibiotic prophylaxis) and patient outcomes (e.g., surgical wound infection rates). SURVEILLANCE Traditionally, infection-control practitioners have surveyed inpatients for infections acquired in hospitals (defined as those neither present nor incubating at the time of admission). Surveillance involves review of microbiology laboratory results, “shoe-leather” epidemiology on nursing wards, and application of standardized definitions of infection. Some infection-control programs use computerized hospital databases for algorithm-driven electronic surveillance (e.g., of vascular catheter and surgical wound infections). Commercial health care information systems that facilitate these functions are considered “value-added” products. Most hospitals aim surveillance at infections associated with a high level of morbidity or expense. Qualityimprovement activities in infection control have led to increased surveillance of personnel compliance with infection-control policies (e.g., adherence to influenza vaccination recommendations).The growing number of states that require public reporting of processes for prevention of health care–associated infection and/or patient outcomes has added new complexity to what hospitals measure and how they measure it. Results of surveillance are expressed as rates. In general, 5–10% of patients develop nosocomial infections— a rate that, as patient advocates emphasize, has remained unchanged for 20–30 years. However, such broad statistics have little value unless qualified by duration of risk, site of infection, patient population, and exposure to risk factors. Meaningful denominators for infection rates include the number of patients exposed to a specific risk (e.g., patients using mechanical ventilators) or the number of intervention days (e.g., 1000 patient-days on a ventilator).

Temporal trends in rates should be reviewed, and rates should be compared with regional and national benchmarks. However, even comparison rates generated by the National Healthcare Safety Network (NHSN) have not been validated independently and represent a nonrandom sample of hospitals. [NHSN is the successor to the National Nosocomial Infections Surveillance System, a program of the Centers for Disease Control and Prevention (CDC) that collected data from more than 350 hospitals that use standardized definitions of nosocomial infections.] Interhospital comparisons may be misleading because of the wide range in risk factors and severity of underlying illnesses. Although systems for making adjustments for these factors either are rudimentary or have not been well validated, process measures (e.g., adherence to hand hygiene) do not usually require risk adjustment, and outcome measures (e.g., cardiac surgery wound infection rates) can identify hospitals with higher infection rates (e.g., in the top quartile) for further evaluation. Moreover, temporal analysis of an individual hospital’s process and infection outcome rates helps to determine whether control measures are succeeding and where increased efforts should be focused.


Nosocomial infections follow basic epidemiologic patterns that can help to direct prevention and control measures. Nosocomial pathogens have reservoirs, are transmitted by predictable routes, and require susceptible hosts. Reservoirs and sources exist in the inanimate environment (e.g., tap water contaminated with Legionella) and in the animate environment (e.g., infected or colonized health care workers, patients, and hospital visitors).The mode of transmission usually is either cross-infection (e.g., indirect spread of pathogens from one patient to another on the inadequately cleaned hands of hospital personnel) or autoinoculation (e.g., aspiration of oropharyngeal flora into the lung along an endotracheal tube). Occasionally, pathogens (e.g., group A streptococci and many respiratory viruses) are spread from person to person via infectious droplets released by coughing or sneezing. Much less common—but often devastating in terms of epidemic risk—is true airborne spread of droplet nuclei (as in nosocomial chickenpox) or common-source spread by contaminated materials (e.g., contaminated IV fluids). Factors that increase host susceptibility include underlying conditions and the many medical–surgical interventions and procedures that bypass or compromise normal host defenses. Through their programs, hospitals’ infection-control committees must determine general and specific control measures. Given the prominence of cross-infection, hand hygiene is the single most important preventive measure in hospitals. Health care workers’ rates of adherence to hand-hygiene recommendations are abysmally low (<50%). Reasons cited include inconvenience, time pressures, and skin damage from frequent washing.

Sinkless alcohol rubs are quick and highly effective and actually improve hand condition since they contain emollients and allow the retention of natural protective oils that would be removed with repeated rinsing. Use of alcohol hand rubs between patient contacts is now recommended for all health care workers except when the hands are visibly soiled, in which case washing with soap and water is still required. improved asepsis in handling and earlier removal of invasive devices, but the maintenance of such gains requires ongoing efforts. It is especially noteworthy that turnover or shortages of trained personnel jeopardize safe and effective patient care and have been associated with increased infection rates. Urinary Tract Infections Urinary tract infections (UTIs) account for as many as 40–45% of nosocomial infections; up to 3% of bacteriuric patients develop bacteremia. Although UTIs contribute only 10–15% to prolongation of hospital stay and to extra costs, these infections are important reservoirs and sources for spread of antibiotic-resistant bacteria in hospitals. Almost all nosocomial UTIs are associated with preceding instrumentation or indwelling bladder catheters, which create a 3–10% risk of infection each day. UTIs generally are caused by pathogens that spread up the periurethral space from the patient’s perineum or gastrointestinal tract—the most common pathogenesis in women—or via intraluminal contamination of urinary catheters, usually due to cross-infection by caregivers who are irrigating catheters or emptying drainage bags. Pathogens come occasionally from inadequately disinfected urologic equipment and rarely from contaminated supplies.

Hospitals should closely monitor essential performance measures for preventing nosocomial UTIs (Table 13-2). Sealed catheter–drainage tube junctions can help to prevent breaks in the system. Approaches to the prevention of UTIs also have included use of topical meatal antimicrobials, drainage bag disinfectants, and anti-infective catheters. None of the latter three measures is considered routine. In fact, a recent meta-analysis suggests that silver alloy–coated anti-infective catheters do not reduce the incidence of bacteriuria from that occurring with silicone catheters. Administration of systemic antimicrobial agents for other purposes decreases the risk of UTI during the first 4 days of catheterization, after which resistant bacteria or yeasts emerge as pathogens. Selective decontamination of the gut is also associated with a reduced risk. Again, however, neither approach is routine. Irrigation of catheters, with or without antimicrobial agents, may actually increase the risk of infection.A condom catheter for men without bladder obstruction may be more acceptable than an indwelling catheter, but the infection risks with the two types are similar unless the condom catheter is carefully maintained.The role of suprapubic catheters in preventing infection is not well defined. Treatment of UTIs is based on the results of quantitative urine cultures (Chap. 27).

The most common pathogens are Escherichia coli, nosocomial gram-negative bacilli, enterococci, and Candida. Several caveats apply in the treatment of institutionally acquired infection. First, in patients with chronic indwelling bladder catheters, especially those in long-term care facilities, “catheter flora”—microorganisms living on encrustations within the catheter lumen—may differ from actual urinary tract pathogens.Therefore, for suspected infection in the setting of chronic catheterization (especially in women), it is useful to replace the bladder catheter and to obtain a freshly voided urine specimen. Second, as in all nosocomial infections, at the time treatment is initiated on the basis of a positive culture, it is useful to repeat the culture to verify the persistence of infection.Third, the frequency with which UTIs occur may lead to the erroneous assumption that this site alone is the source of infection in a febrile hospitalized patient. Fourth, recovery of Staphylococcus aureus from urine cultures may result from hematogenous seeding and may indicate an occult systemic infection. Finally, although Candida is now the most common pathogen in nosocomial UTIs in patients on intensive care units (ICUs), treatment of candiduria is often unsuccessful and is recommended only when there is upper-pole invasion, obstruction, neutropenia, or immunosuppression. Pneumonia Pneumonia accounts for 15–20% of nosocomial infections but has been responsible for 24% of extra hospital days and 39% of extra costs—i.e., 6 days and the associated costs per episode. Almost all cases of bacterial nosocomial pneumonia are caused by aspiration of endogenous or hospital-acquired oropharyngeal (and occasionally gastric) flora. Nosocomial pneumonias are associated with more deaths than are infections at any other body site. However, attributable mortality for ventilator-associated pneumonia—the most common and lethal form of nosocomial pneumonia—is in the 6–14% range; this figure suggests that the risk of dying from nosocomial pneumonia is affected greatly by other factors, including comorbidities, inadequate antibiotic treatment, and the involvement of specific pathogens (particularly Pseudomonas aeruginosa and Acinetobacter). Surveillance and accurate diagnosis of pneumonia are often problematic in hospitals because many patients, especially those in the ICU, have abnormal chest roentgenographs, fever, and leukocytosis potentially attributable to multiple causes.

Viral pneumonias, which are particularly important in pediatric and immunocompromised patients, are discussed in the virology section and in Chap. 17. Risk factors for nosocomial pneumonia, particularly ventilator-associated pneumonia, include those events that increase colonization by potential pathogens (e.g., prior antimicrobial therapy, contaminated ventilator circuits or equipment, or decreased gastric acidity), those that facilitate aspiration of oropharyngeal contents into the lower respiratory tract (e.g., intubation, decreased levels of consciousness, or presence of a nasogastric tube), and those that reduce host defense mechanisms in the lung and permit overgrowth of aspirated pathogens (e.g., chronic obstructive pulmonary disease, old age, or upper abdominal surgery). Control measures for pneumonia (Table 13-2) are aimed at the remediation of risk factors in general patient care (e.g., minimizing aspiration-prone supine positioning) and at meticulous aseptic care of respirator equipment (e.g., disinfecting or sterilizing all inline reusable components such as nebulizers, replacing tubing circuits at intervals of >48 h—rather than more frequently—to lessen the number of breaks in the system, and teaching aseptic technique for suctioning). The benefits of selective decontamination of the oropharynx and gut with nonabsorbable antimicrobial agents and/or use of shortcourse postintubation systemic antibiotics have been controversial.

Among the logical preventive measures that require further investigation are the use of endotracheal tubes that provide channels for subglottic drainage of secretions and the use of noninvasive mechanical ventilation whenever feasible. It is noteworthy that reducing the rate of ventilator-associated pneumonia most often has not reduced overall ICU mortality; this fact suggests that this infection is a marker for patients with an otherwiseheightened risk of death. The most likely pathogens for nosocomial pneumonia and treatment options are discussed in Chap. 17. Several considerations regarding diagnosis and treatment are worth emphasizing. Clinical criteria for diagnosis (e.g., fever, leukocytosis, development of purulent secretions, new or changing radiographic infiltrates, changes in oxygen requirement or ventilator settings) have high sensitivity but relatively low specificity. These criteria are most useful for selecting patients for bronchoscopic or nonbronchoscopic procedures that yield lower respiratory tract samples protected from upper-tract contamination; quantitative cultures of such specimens have diagnostic sensitivities in the range of 80%. Early-onset nosocomial pneumonia, which manifests within the first 4 days of hospitalization, is most often caused by communityacquired pathogens, such as Streptococcus pneumoniae and Haemophilus species.

Late-onset pneumonias most commonly are due to S. aureus, P. aeruginosa, Enterobacter species, Klebsiella pneumoniae, or Acinetobacter—a pathogen of increasing concern in many ICUs. When invasive techniques are used to diagnose ventilator-associated pneumonia, the proportion of isolates accounted for by gram-negative bacilli decreases from 50–70% to 35–45%. Infection is polymicrobial in as many as 20–40% of cases. The role of anaerobic bacteria in ventilator-associated pneumonia is not well defined. A recent study suggested that 8 days is an appropriate duration of therapy for nosocomial pneumonia, with a longer duration (15 days in this study) when the pathogen is Acinetobacter or P. aeruginosa. Finally, in febrile patients (particularly those who have endotracheal and/or nasogastric tubes), more occult sources of respiratory tract infection, especially bacterial sinusitis and otitis media, should be considered. Surgical Wound Infections Wound infections account for up to 20–30% of nosocomial infections but contribute up to 57% of extra hospital days and 42% of extra costs. The average wound infection has an incubation period of 5–7 days (longer than many postoperative stays), and many procedures are now performed on an outpatient basis.Thus the incidence of wound infections has become difficult to assess. These infections usually are caused by the patient’s endogenous or hospital-acquired skin and mucosal flora and occasionally are due to airborne spread of skin squames that may be shed into the wound from members of the operating-room team.True airborne spread of infection through droplet nuclei is rare in operating rooms unless there is a “disseminator” (e.g., of group A streptococci or staphylococci) among the staff. In general, the most common risks for postoperative wound infection are related to the surgeon’s technical skill, the patient’s underlying diseases (e.g., diabetes mellitus, obesity) or advanced age, and inappropriate timing of antibiotic prophylaxis. Additional risk factors include the presence of drains, prolonged preoperative hospital stays, shaving of the operative site by razor the day before surgery, a long duration of surgery, and infection at remote sites (e.g., untreated UTI).

The substantial literature related to risk factors for surgical-site infections and the recognized morbidity and cost of these infections have led to national prevention efforts—the Surgical Infection Prevention (SIP) Project, the Institute for Healthcare Improvement (IHI) 100,000 Lives Campaign, and the Surgical Care Improvement Project (SCIP)—and to recommendations for “bundling” of evidence-based preventive measures (Table 13-2). Additional measures include attention to technical surgical issues and operating-room asepsis (e.g., avoiding open or prophylactic drains) and preoperative therapy for active infection. Reporting of surveillance results to surgeons has been associated with reductions in infection rates. The use of preoperative intranasal mupirocin to eliminate that reservoir for S. aureus, preoperative antiseptic bathing, and supplemental intra- and postoperative oxygen remain controversial because of conflicting study results. The increasingly extensive review of infection rates by regulatory agencies and third-party payers emphasizes the importance of stratifying rates by patient-related risk factors and of developing meaningful systems for wound surveillance after the patient’s discharge from the hospital or clinic (when >50% of infections first become apparent) or for use of surrogate markers of wound infection (e.g., prolonged postoperative antibiotic courses). The epidemic of mad cow disease, centered in the United Kingdom, and associated human cases of variant Creutzfeldt-Jakob disease (Chap. 101) caused by disinfection-resistant prion agents have led to revised recommendations for decontaminating surgical instruments, especially those used for operations on the central nervous system or in patients with dementing illness of unknown etiology.

The process of diagnosing and treating wound infections begins with a careful assessment of the surgical site in the febrile postoperative patient. Clinical findings range from obvious cellulitis or abscess formation to subtler clues, such as a sternal “click” after open heart surgery. Diagnosis of deeper organ-space infections or subphrenic abscesses requires a high index of suspicion and the use of CT or MRI. Diagnosis of infections of prosthetic devices, such as orthopedic implants, may be particularly difficult and often requires the use of interventional radiographic techniques to obtain periprosthetic specimens for culture. The most common pathogens in postoperative wound infections are S. aureus, coagulase-negative staphylococci, and enteric and anaerobic bacteria. In rapidly progressing postoperative infections, which manifest within 24–48 h of a surgical procedure, the level of suspicion regarding group A streptococcal or clostridial infection (Chaps. 36 and 42) should be high.Treatment of postoperative wound infections requires drainage or surgical excision of infected or necrotic material and antibiotic therapy aimed at the most likely or laboratory-confirmed pathogens. Infections Related to Vascular Access and Monitoring Intravascular devices are common causes of local site infection and cause up to 50% of nosocomial bacteremias; central vascular catheters (CVCs) account for 80–90% of these infections. National estimates indicate that as many as 200,000 bloodstream infections associated with CVCs occur each year in the United States, with an attributable mortality of 12–25% and an estimated cost of $25,000 per episode; one-third to one-half of these episodes occur in ICUs.With increasing care of seriously ill patients in the community, vascular catheter–associated bloodstream infections acquired in outpatient settings may become as frequent as those acquired in hospitals. This possibility emphasizes the need to broaden surveillance activities. Catheter-related bloodstream infections derive largely from the cutaneous microflora of the insertion site, with pathogens migrating extraluminally to the catheter tip, usually during the first week after insertion. In addition, contamination of hubs of CVCs or of the ports of “needleless” systems may lead to intraluminal infection over longer periods, particularly with surgically implanted or cuffed catheters. Intrinsic contamination of infusate, although rare, is the most common cause of epidemic device-related bloodstream infection; extrinsic contamination may cause up to half of endemic bacteremias related to arterial infusions used for hemodynamic monitoring.

The most common pathogens isolated from vascular device–associated bacteremias include coagulasenegative staphylococci, S. aureus (with up to 50% or more of isolates in the United States resistant to methicillin), enterococci, nosocomial gram-negative bacilli, and Candida. Many pathogens, especially staphylococci, produce extracellular polysaccharide biofilms that facilitate attachment to catheters and provide sanctuary from antimicrobial agents.“Quorum-sensing” proteins help bacterial cells communicate during biofilm development. Infections related to vascular catheters and monitoring devices may be the most preventable of nosocomial infections. Evidence-based bundles of control measures (Table 13-2) have been strikingly effective, eliminating all infections in one ICU study. Hospitals should periodically monitor adherence to these performance indicators. Use of antimicrobial- or antiseptic-impregnated CVCs does not appear necessary if the prevention bundle is fully implemented. Additional control measures for infections associated with vascular access include using a chlorhexidine-impregnated patch at the skin-catheter junction; avoiding the femoral site for catheterization because of higher risk of infection (most likely related to the density of the skin flora); moving peripheral catheters to a new site at specified intervals (e.g., every 72–96 h), which may be facilitated by use of an IV therapy team; and applying disposable transducers for pressure monitoring and aseptic technique for accessing transducers or other vascular ports. Improvements in composition of semitransparent access-site dressings and potential nursing benefits (ease of bathing and site inspection, protection of site from secretions) favor the use of such coverings. Unresolved issues include the best frequency for rotation of CVC sites (given that guidewire-assisted catheter changes at the same site do not lessen and may even increase infection risk); the appropriate role of mupirocin ointment, a topical antibiotic with excellent antistaphylococcal activity, in site care; the relative degrees of risk posed by peripherally inserted central catheters (PICC) lines; and the risk-benefit of prophylactic use of heparin (to avoid catheter thrombi, which may be associated with increased risk of infection) or of vancomycin or alcohol (as catheter flushes or “locks”—i.e., concentrated anti-infective solutions instilled into the catheter lumen) for high-risk patients.

Vascular device–related infection is suspected on the basis of the appearance of the catheter site or the presence of fever or bacteremia without another source in patients with vascular catheters. The diagnosis is confirmed by the recovery of the same species of microorganism from peripheral-blood cultures (preferably two cultures drawn from peripheral veins by separate venipunctures) and from semiquantitative or quantitative cultures of the vascular catheter tip. Less commonly used diagnostic measures include differential time to positivity (>2 h) for blood drawn through the vascular access device compared with a sample from a peripheral vein or differences in quantitative cultures (a 5- to 10-fold or greater “step-up”) for blood samples drawn simultaneously from a peripheral vein and from a CVC.When infusionrelated sepsis is considered (e.g., because of the abrupt onset of fever or shock temporally related to infusion therapy), a sample of the infusate or blood product should be retained for culture. Therapy for vascular access–related infection is directed at the pathogen recovered from the blood and/or infected site. Important considerations in treatment are the need for an echocardiogram (to evaluate the patient for bacterial endocarditis), the duration of therapy, and the need to remove potentially infected catheters. In one report, approximately one-fourth of patients with intravascular catheter–associated S. aureus bacteremia who were studied by transesophageal echocardiography had evidence of endocarditis; this test may be useful in determining the appropriate duration of treatment. Detailed consensus guidelines for the management of intravascular catheter–related infections have been published and recommend catheter removal in most cases of bacteremia or fungemia due to nontunneled CVCs.When attempting to salvage a potentially infected catheter, some clinicians use the “antibiotic lock” technique, which may facilitate penetration of infected biofilms, in addition to systemic antimicrobial therapy. In one study of hemodialysis catheters, only about one-third of salvage attempts were successful, although delayed removal did not appear to increase the risk of complications. Often, a potentially infected CVC may be exchanged over a guidewire. If cultures of the removed catheter tip are positive, the replacement catheter will be moved to a new site; if the tip cultures are negative, the replacement catheter may remain in the original site but may be at increased risk of subsequent infection due to this manipulation.

The authors of the consensus guidelines advise that the decision to remove a tunneled catheter or implanted device suspected of being the source of bacteremia or fungemia should be based on the severity of the patient’s illness, the strength of the evidence that the device is infected, an assessment of the specific pathogens, and the presence of local or systemic complications. For patients with track-site infection, successful therapy without catheter removal is unusual. For patients with suppurative venous thrombophlebitis, excision of the affected vein is usually required.


Written policies for the isolation of infectious patients are a standard component of infection-control programs. In 1996, the CDC revised its isolation guidelines to make them simpler; to recognize the importance of all body fluids, secretions, and excretions in the transmission of nosocomial pathogens; and to focus precautions on the major routes of infection transmission. These policies are currently being updated by the CDC to include integrated guidelines for control of multidrug-resistant organisms. Standard precautions are designed for the care of all patients in hospitals and aim to reduce the risk of transmission of microorganisms from both recognized and unrecognized sources. These precautions include gloving as well as hand cleansing for potential contact with (1) blood; (2) all other body fluids, secretions, and excretions, whether or not they contain visible blood; (3) nonintact skin; and (4) mucous membranes.

Depending on exposure risks, standard precautions also include use of masks, eye protection, and gowns. Precautions for the care of patients with potentially contagious clinical syndromes (e.g., acute diarrhea) or with suspected or diagnosed colonization or infection with transmissible pathogens are based on probable routes of transmission: airborne, droplet, and contact. Sets of precautions may be combined for diseases that have more than one route of transmission (e.g., varicella). Because some prevalent antibiotic-resistant pathogens, particularly vancomycin-resistant enterococci (VRE), may be present on intact skin of patients in hospitals, some experts recommend gloving for all contact with patients who are acutely ill and/or from high-risk units, such as ICUs.Wearing gloves does not replace the need for hand hygiene because hands occasionally become contaminated during wearing or removal of gloves. Some studies have suggested that use of gowns and gloves compared with routine care of patients (i.e., using neither of these barriers) decreases the risk of nosocomial infection; however, the benefit of gowning by personnel beyond that conferred by gloving and hand hygiene is controversial. Nevertheless, requiring increased precaution levels can improve the compliance of health care workers with isolation recommendations by 30%.


Outbreaks and emerging pathogens are always big news but probably account for <5% of nosocomial infections. Concern about emerging pathogens often prompts authorities to require hospitals to develop contingency and response plans.The investigation and control of nosocomial epidemics require that infection-control personnel develop a case definition, confirm that an outbreak really exists (since many apparent epidemics are actually pseudooutbreaks due to surveillance or laboratory artifacts), review aseptic practices and disinfectant use, determine the extent of the outbreak, perform an epidemiologic investigation to determine modes of transmission, work closely with microbiology personnel to culture for common sources or personnel carriers as appropriate and to type epidemiologically important isolates, and heighten surveillance to judge the effect of control measures. Control measures generally include reinforcing routine aseptic practices and hand hygiene during a search for compliance problems that may have fostered the outbreak, ensuring appropriate isolation of cases (and instituting cohort isolation and nursing if needed), and implementing further controls on the basis of the investigation’s findings. Examples of some emerging and potential epidemic problems follow.

Viral Respiratory Infections: SARS and Influenza Infections caused by the severe acute respiratory syndrome (SARS)–associated coronavirus challenged health care systems globally in 2003 (Chap. 87). Basic infection- control measures helped to keep the worldwide case and death counts at ∼8000 and ∼800, respectively, although SARS was unforgiving of lapses in protocol adherence or laboratory biosafety.The epidemiology of SARS—spread largely in households once patients were ill or in hospitals—contrasts markedly with that of influenza (Chap. 88), which is often contagious a day before symptom onset, can spread rapidly in the community among nonimmune persons, and kills as many as 30,000 persons each year in the United States. Control of influenza has depended on (1) the use of effective vaccines, with increasingly broad recommendations for vaccination and emphasis on vaccination of health care workers; (2) the use of antiviral medications for early treatment and for prophylaxis as part of outbreak control, especially in high- risk settings like nursing homes or hospitals; and (3) infection control (surveillance and droplet precautions) for symptomatic patients. Concerns about avian (H5N1) and pandemic influenza have led to recommendations for “respiratory hygiene and cough etiquette” and “source containment” (e.g., use of face masks and spatial separation) for outpatients with potentially infectious respiratory illnesses; to the concept of “social distancing” (e.g., closing community venues such as shopping malls) in the event of a pandemic; and to debate about the level of avian influenza respiratory protection required for health care workers—i.e., whether to use the higher-efficiency N95 respirators recommended for airborne isolation rather than the surgical masks used for droplet precautions.

Nosocomial Diarrhea A new, more virulent strain of Clostridium difficile has emerged in North America, and overall rates of C. difficile– associated diarrhea (Chap. 43) have increased in U.S. hospitals during the past few years. The potential role of exposure to newer fluoroquinolone antibiotics in driving these changes is being investigated. C. difficile control measures include judicious use of all antibiotics; heightened suspicion for “atypical” presentations (e.g., toxic megacolon or leukemoid reaction without diarrhea); and early diagnosis, treatment, and contact precautions. Outbreaks of norovirus infection (Chap. 91) in U.S. and European health care facilities appear to be increasing in frequency, with the virus often introduced by ill visitors or health care workers. This pathogen should be suspected when nausea and vomiting are prominent aspects of bacterial culture–negative diarrheal syndromes. Contact precautions may need to be augmented by aggressive environmental cleaning (given the persistence of norovirus on inanimate objects) and active exclusion of ill staff and visitors.

Chickenpox Infection-control practitioners institute a varicella exposure investigation and control plan whenever health care workers either (1) are exposed to chickenpox (Chap. 81) in the community or through patients with initially unrecognized infections or (2) work during the 24 h before developing chickenpox. The names of exposed workers and patients are obtained; medical histories are reviewed, and (if necessary) serologic tests for immunity are conducted; physicians are notified of susceptible exposed patients; postexposure prophylaxis with varicella-zoster immune globulin (VZIG) is considered for immunocompromised or pregnant contacts (see Table 81-1); varicella vaccine is recommended or preemptive use of acyclovir is considered as an alternative strategy in other susceptible persons; and susceptible exposed employees are furloughed during the at-risk period for disease (8–21 days or—if VZIG has been administered—28 days). Routine varicella vaccination of children and susceptible employees can markedly decrease risk and frequency of exposures. Tuberculosis Important measures for the control of tuberculosis (Chap. 66) include prompt recognition, isolation, and treatment of cases; recognition of atypical presentations (e.g., lower-lobe infiltrates without cavitation); use of negative-pressure, 100% exhaust, private isolation rooms with closed doors and 6–12 or more air changes per hour; use of N95 “respirators” (approved by the National Institute for Occupational Safety and Health) by caregivers entering isolation rooms; possible use of highefficiency particulate air filter units and/or ultraviolet lights for disinfecting air when other engineering controls are not feasible or reliable; and follow-up skin-testing of susceptible personnel who have been exposed to infectious patients before isolation. The use of new serologic tests, rather than skin tests, in the diagnosis of latent tuberculosis for infection control purposes is being studied.

Group A Streptococcal Infections The potential for an outbreak of group A streptococcal infection (Chap. 36) should be considered when even a single nosocomial case occurs. Most outbreaks involve surgical wounds and are due to the presence of an asymptomatic carrier in the operating room. Investigation can be confounded by carriage at extrapharyngeal sites such as the rectum and vagina. Health care workers in whom carriage has been linked to nosocomial transmission of group A streptococci are removed from the patient-care setting and are not permitted to return until carriage has been eliminated by antimicrobial therapy. Fungal Infections Fungal spores are common in the environment, particularly on dusty surfaces. When dusty areas are disturbed during hospital repairs or renovation, the spores become airborne. Inhalation of spores by immunosuppressed (especially neutropenic) patients creates a risk of pulmonary and/or paranasal sinus infection and disseminated aspergillosis (Chap. 108). Routine surveillance among neutropenic patients for infections with filamentous fungi, such as Aspergillus and Fusarium, helps hospitals to determine whether they are facing unduly extensive environmental risks.As a matter of routine, hospitals should inspect and clean air-handling equipment, review all planned renovations with infection-control personnel and subsequently construct appropriate barriers, remove immunosuppressed patients from renovation sites, and consider the use of high-efficiency particulate air intake filters for rooms housing immunosuppressed patients.

Legionellosis Nosocomial Legionella pneumonia (Chap. 49) is most often due to contamination of potable water and predominantly affects immunosuppressed patients, particularly those receiving glucocorticoid medication.The risk varies greatly within and among geographic regions, depending on the extent of hospital hot-water contamination and on specific hospital practices (e.g., inappropriate use of nonsterile water in respiratory therapy equipment). Laboratory-based surveillance for nosocomial Legionella should be performed, and a diagnosis of legionellosis should probably be considered more often than it is. If cases are detected, environmental samples (e.g., tap water) should be cultured. If cultures yield Legionella and if typing of clinical and environmental isolates reveals a correlation, eradication measures should be pursued. An alternative approach is to periodically culture tap water in wards housing high-risk patients. If Legionella is found, a concerted effort should be made to culture samples from all patients with nosocomial pneumonia for Legionella. Antibiotic-Resistant Bacteria Control of antibiotic resistance, particularly in outbreaks (Table 13-3), depends on close laboratory surveillance, with early detection of problems; on aggressive reinforcement of routine asepsis (e.g., hand hygiene); on implementation of barrier precautions for all colonized and/or infected patients; on use of patient-surveillance cultures to more fully ascertain the extent of patient colonization; and on timely initiation of an epidemiologic investigation when rates increase.

Colonized personnel who are implicated in nosocomial transmission and patients who pose a threat may be decontaminated. In a few ICUs, selective decontamination of patients has been used successfully as a temporary emergency control measure for outbreaks of infection due to gram-negative bacilli. Other promising ICU control measures include daily bathing of patients with chlorhexidine and enforcement of environmental cleaning; in recent trials, each of these measures reduced cross-transmission of VRE. The value of “search-and-destroy” methods—i.e., the use of active surveillance cultures to detect and isolate the “resistance iceberg” of patients colonized with methicillin-resistant S. aureus (MRSA) or VRE—in non-outbreak settings has been controversial but is credited with elimination of nosocomial MRSA in the Netherlands and Denmark. Currently, several antibiotic resistance problems are of particular health care concern. First, the emergence of community-acquired MRSA has been dramatic in many countries, with as many as 50% of community-acquired “staph infections” in some U.S. cities now caused by strains resistant to â-lactam antibiotics (Chap. 35). The potential incursion of these strains into hospitals and the resulting impact on control of nosocomial MRSA infections are of enormous concern. Second, in the ongoing global reemergence of nosocomial multidrug-resistant gram-negative bacilli, new problems include plasmid-mediated resistance to fluoroquinolones, metallo-â-lactamase-mediated resistance to carbapenems, and panresistant strains of Acinetobacter. Many of these multidrug-resistant strains are susceptible only to colistin, which has led to a “rediscovery” and renewed use of this drug. Finally, clinical infections with MRSA strains exhibiting high-level vancomycin resistance due to VREderived plasmids have been reported in several patients in the United States, often in the setting of prolonged or repeated treatment with vancomycin and/or VRE colonization. The detection of any of these current problems should trigger an epidemiologic investigation and aggressive infection-control measures. Because the excessive use of broad-spectrum antibiotics underlies many resistance problems, aggressive antibioticcontrol policies must be considered a cornerstone of resistance-control efforts. Recommendations for “antibiotic stewardship” are being promulgated by the Infectious Diseases Society of America.

Although the efficacy of antibiotic-control measures in reducing rates of antimicrobial resistance has not been proven in prospective controlled trials, it seems worthwhile to restrict the use of particular agents to narrowly defined indications in order to limit selective pressure on the nosocomial flora. Bioterrorism and Other “Surge-Event” Preparedness The horrific attack on the World Trade Center in New York City on September 11, 2001; the subsequent mailings of anthrax spores in the United States; and recently exposed terrorist plans and activities in the United Kingdom and elsewhere have made bioterrorism a prominent source of concern to hospital infection-control programs. The essentials for hospital preparedness (Table 13-4) entail education, internal and external communication, and risk assessment. Up-to-date information on a variety of bioterrorism-associated issues is available from the CDC (


An institution’s employee health service is a critical component of its infection-control efforts. New employees should be processed through the service, where a contagious- disease history can be taken; evidence of immunity to a variety of diseases, such as hepatitis B, chickenpox, measles, mumps, and rubella, can be sought; immunizations for hepatitis B, measles, mumps, rubella, and varicella can be given as needed, and a reminder about the need for yearly influenza immunization can be imparted; baseline and “booster” purified protein derivative of tuberculin skin-testing or serologic testing for tuberculosis can be performed; and education about personal responsibility for infection control can be initiated. Evaluations of employees should be codified to meet the requirements of accrediting and regulatory agencies. The employee health service must have protocols for dealing with workers who have been exposed to contagious diseases, such as those percutaneously or mucosally exposed to the blood of patients infected with HIV or hepatitis B or C virus. For example, postexposure HIV prophylaxis with a combination of two or three antiretroviral agents is recommended; free consultation is available from the CDC PEPLine (888-HIV-4911). Protocols are also needed for dealing with caregivers who have common contagious diseases (such as chickenpox, group A streptococcal infections, respiratory infections, and infectious diarrhea) and for those who have less common but high-visibility public health problems (such as chronic hepatitis B or C or HIV infection) for which exposurecontrol guidelines have been published by the CDC and by the Society for Healthcare Epidemiology of America.