key: cord-0008486-h1gnp62r
authors: Neu, Harold C
title: Unusual nosocomial infections
date: 2004-06-29
journal: Dis Mon
DOI: 10.1016/0011-5029(84)90018-x
sha: 9720bb0a2e4f97f861926e30d14357ea179fb46b
doc_id: 8486
cord_uid: h1gnp62r

nan

IN THE PAST few decades there has been striking progress in many areas of medical science. Many patients who in the 1950s and in the early part of the 1960s would have died of cardiovascular, pulmonary, hematologic, gastrointestinal, or oncologic problems can be saved through the use of new pharmacologic agents or a variety of life support devices. This progress in medicine has at the same time added a new dimension to the problem of hospital-acquired, or nosocomial, infection.

The word nosocomial was first used in 1940 by Joyce Wright in a paper discussing streptococcal infection in children's wards in London. 1 Nosocomial infections in the past were defined as those diagnosed within 48 hours after admission. It may be more proper to define nosocomial infections as those that are neither present nor incubating at the time of admission, since it is now recognized that a number of infections such as hepatitis, sternal wound infections, or fungal infections may be acquired in the hospital yet may not be recognized until long incubation periods have passed. Data presented by Allen et al. 2 at the Second International Conference on Nosocomial Infections, held in August 1980, indicated that the nosocomial infection rate, which was 341 per 10,000 patients discharged in 1975, had declined to 329 per 10,000 patients in 1979. Infection rates for community-teaching and municipal hospitals declined from 1970 to 1979, whereas those for community and university hospitals did not decline. There has been in the past decade, and continuing to the present, a decline in the rate of surgical wound infection, whereas infections on other services have increased. Of interest, there was an increase in bacteremia in the decade of the 1970s. Observations at our own institution suggest that bacteremia has continued to be a significant problem, and that other forms of infection have remained at a level similar to the level of the previous decade, despite improvements in infection control within the hospital.

The distribution of endemic nosocomial infections occurring in the hospital has remained fairly stable for the past decade. Urinary tract infections account for 35%-40% of infections, surgical wounds for 25%-30%, pneumonia for 15%-20%, skin structure infections for 5%-10%, bacter-emia for 5%, and other forms of infection for about 5%-10%. Although great progress has been made in the past two decades in improving urinary catheter care with the institution of closed drainage systems, urinary tract infections related to instrumentation of the urinary bladder continue to be a major kind of nosocomial infection. 3 The same may be said for respiratory tract infections. 4 There has been a marked increase in the use of respiratory support systems in patients who would have succumbed to pulmonary insufficiency in previous years. Although problems associated with contaminated respiratory care equipment are distinctly less frequent today than they were in the 1960s, nonetheless, nosocomial pneumonitis remains a major cause of death in hospitalized patients. ~' 6 The pathogens causing nosocomial infections have not significantly changed in the past decade. Escherichia coli is still the most frequent organism isolated in approximately 20% of infections with Staphylococcus aureus and enterococci both being encountered in about 10% of infections, followed by Pseudomonas aeruginosa in 9% to 10% and Klebsiella infections and Proteus infections in about 7% to 8%. The frequency of Klebsiella infections and Proteus infections has declined to some extent in recent years, whereas there has been an increase in the number of infections caused by Enterobacter species and Serratia marcescens. One change in the organisms causing infections has been in the antimicrobial susceptibilities. There has been an increase in resistance to first-generation cephalosporins among organisms such as Klebsiella and E. coli, and organisms such as Serratia and Enterobacter have shown resistance both to aminoglycosides and to fourth-$eneration penicillins and third-generation cephalosporins. ~ Analysis of epidemics of nosocomial infections is difficult since epidemics usually involve infection by unusual organisms or particularly resistant organisms. Stamm et al. s noted that epidemics accounted for a small proportion of preventable infections acquired in the hospital, but epidemics have been extremely useful in defining the modes of spread, the source, and the methods to prevent and to control nosocomial infections. The pathogens involved in epidemics have been primarily S. aureus, Klebsiella, Salmonella, hepatitis B, Pseudomonas, and group A streptococci (Table 1) . Although urinary tract infections are a common form of nosocomial epidemic infections, gastroenteritis, skin infections, bacteremia, meningitis, and hepatitis are more frequently involved, since these are the infections for which the aid of the Centers for Disease Control (CDC) has been sought to solve the problem of infectious epidemics.

Certain forms of nosocomial infection occur more frequently in immunocompromised patients. Patients with underlying disease such as acute leukemia, Hodgkin's disease, non-Hodgkin's lymphoma, and other tumors and who have renal transplants are particularly at risk for infection by unusual viruses, fungi, and parasites. Nonetheless, these patients develop infection at the same sites as do other hospitalized patients, and by the most common organisms.

This monograph discusses infections due to some of the less common organisms and particularly species that are emerging as more important with our ability to recognize unusual pathogens ( Table 2) . Some of these pathogens may be ordinary microorganisms with unusual antimicrobial resistance patterns which have been selected as the result of extensive antimicrobial use, similar to the infections due to Staphylococcus aureus 10 12 Enterococcus sp.

Pseudomonas sp. 9 4 Proteus sp.

KlebsieUa sp. Klebsiella and Pseudomonas following the introduction of penicillin and the cephalosporins. Or they may be more fastidious organisms whose slow growth has rendered recognition difficult.

A CINETOBA CTER Acinetobacter calcoaceticus var. anitratus or var. lwoffi is a nonfermentative gram-negative coccal bacillary organism that may form part of the normal flora of the skin, conjunctiva, and perineal area. Acinetobacter has been considered to be a relatively avirulent organism, although it has definite pathogenic potential. It has been associated with a wide variety of illnesses in debilitated, hospitalized patients. 9 RespiratSry tract infection in intensive care patients has been the most common problem associated with Acinetobacter infection. Buxton et al., 1° investigating an outbreak in Connecticut, found that one third of hospital personnel in an intensive care unit had transient hand colonization with multiple strains of Acinetobacter. Pharyngeal, vaginal, and rectal carriage was rare, however. Gram-negative bacilli generally are a transient rather than a resident flora on the hands of most individuals. In epidemic periods, hands of personnel may transmit gramnegative bacilli from patient to patient, but most bacteria persist for less than 24 hours on the hands. In the presence of dermatologic disease, however, organisms can persist for a much longer period. In the outbreak studied by Buxton et al., dermatitis involving the hands of a respiratory therapist probably predisposed him to persistent colonization of Acinetobacter. The personnel and patients remained a reservoir of Acinetobacter, causing continued cross-contamination of respiratory therapy equipment and infection of intubated patients in the intensive care unit. In another outbreak studied by Cunha et al., 11 Acinetobacter pulmonary infections were traced to Wright respirometers. Over a 30-day period, Acinetobacter accounted for hospital-acquired pneumonia in ten patients and colonized the upper respiratory tract in an additional nine patients. The Wright respirometer acted as a means of aerosolizing fluid into the lung.

In addition to respiratory tract infections, bacteremia has also been associated with Acinetobacter. Ramphal and Kluge 9 reported on 13 patients who acquired the organism from the hospital environment and who then developed bacteremia. Conversely, pseudobacteremia with false positive blood cultures has also occurred with Acinetobacter. Syndman et al. 12 reported that children in mist tents had blood cultures positive for Acinetobacter, probably as the result of contamination of the hands of respiratory therapy technicians and blood-drawing personnel. However, children did have colonization of the skin and nose with the organism, and hence would be at risk for development of infection. Contamination of materials has also resulted in an outbreak of Acinetobacter infection associated with peritoneal dialysis. Abrutyn et al. 13 found that over a 4-month period, 14 patients had fluid cultures positive for A. calcoaceticus in dialysis drainage. A water bath used to warm bottles of peritoneal dialysate was the reservoir of the bacteria, and the dialysate became contaminated when the prong of the fluid administration set was inserted through the rubber bung on the dialysis models. This outbreak illustrates the potential importance of environmental reservoirs in infections complicating peritoneal dialysis. In several of these Acinetobacter outbreaks a seasonal incidence was noted, with the organism more frequently isolated during summer months. ~t This would be in keeping with its occurrence on the skin of personnel or in water solutions.

The outbreaks of Acinetobacter illustrate the necessity of careful attention to environmental sources of contamination and to personnel as a source of contamination when the laboratory records increased numbers of isolates of this organism. Microbiology laboratories and hospital infection control personnel must be alerted to investigate possible sources of infection caused by this organism when increased numbers of Acinetobacter infection are noted, particularly in respiratory care units.

Acinetobacter is resistant to a large number of antimicrobial agents. Fortunately, it remains susceptible in most sit-uations to the antipseudomonas penicillins, ticarcillin, and the newer ureido penicillins (derivatives of ampicillin that possess the properties of both ampicillin and carbenicillin; e.g., mezlocillin, azlocillin, and piperacillin). However, when strains of Acinetobacter acquire the TEM* plasmid which mediates production of a ~-lactamase that hydrolyzes these penicillins, as may occur in a respiratory care unit, these organisms will be resistant to these penicillins.

Resistance of Acinetobacter to aminoglycosides varies from 15% (to tobramycin) to as high as 30% (to gentamicin). None of the new third-generation cephalosporins is considered optimal therapy for this organism, although in the future a subset will be susceptible to ceftazidime or to imipenem. 15 LEGIONELLA Legionella is an aerobic, fastidious, weakly gram-negative flagellated bacterium that has been recognized as a human pathogen since the 1976 outbreak of Legionnaires' disease in Philadelphia. 16 A number of different Legionella species have now been defined, including L. pneumophila, L. micdadei (Pittsburgh pneumonia agent), L. bozemanni, and L. dumoffi. Legionella is found in the soil and natural water sources. 17 In nature it appears to be associated both • 18 with blue-green algae and with amebae. Although the organisms grow quite well in natural water sources, they also seem to have a great predilection for man-made hot water systems. In the past it was estimated that Legionella accounted for 3%-5% of nosocomial pneumonias, but a more recent study by Muder et al. 19 prospectively evaluating the occurrence of nosocomial pneumonia at a Veterans Administration hospital, where Legionnaires' disease was known to be endemic, and at a community teaching hospital, where Legionnaires' disease had never been documented, demonstrated that at the community hospital 64% of sites in the water distribution system yielded L. pneu-*TEM is the trivial name used to refer to the most common plasmidmediated ~-lactamase. The letters are the initials of the child from whom the first organism carrying the plasmid was isolated. mophila and that Legionnaires' disease accounted for 14.3% of nosocomial pneumonias.

Clinically, L. pneumophila pneumonia and L. micdadei and L. dumoffi infection are similar. Characteristically, the illness is an acute, febrile multisystem illness in which pneumonia is the predominating symptom. There may also be diarrhea, obtundation, hematuria, and impaired liver function. Factors that may suggest that the pneumonia is caused by Legionella are prior corticosteroid medication, lack of symptoms of upper respiratory tract involvement, negative routine bacterial cultures, and failure to respond to therapy with ~-lactam and aminoglycoside antibiotics. Pleuritic chest pain has been noted to be a frequent problem in individuals with Legionnaires' disease, and the course of illness is often complicated by acute renal failure, refractory hypoxemia, and, in some patients, the development of pulmonary abscesses. Initially, however, the illness is difficult to differentiate from a gram-negative pneumonia that has developed in a hospitalized patient.

The case-fatality rates of sporadic and epidemic nosocomial legionellosis range from 25% to 35% in most series. Recently, however, a report from the University of Iowa noted a 46% case-fatality rate in a hematology-oncology unit. e° Variations in case-fatality rates probably are related to differing degrees of immunosuppression and to differences in the severity of the underlying illness in the patient population. Neutropenia in the patients undoubtedly contributes to a higher case-fatality rate. Although most of the early outbreaks of legionellosis were associated with airconditioning apparatus, excavation, or cooling towers, most nosocomial infections now appear to be associated with potable water, el-e4 This has been true for hospitals in the United States, Great Britain, and on the Continent. The precise way Legionella is transmitted to patients is unknown. It is clear, however, that showerheads frequently contain Legionella and can aerosolize Legionella to patients.24, 25 Cordes and colleagues 26 investigated the showerheads in a hospital ward in Chicago where three patients had contracted Legionnaires' disease and found that nine of 16 showerheads were positive for the organism. Each patient had used one of these showers 2-10 days before the onset of symptoms. One problem with these investigations is that showerheads in hospital wards where patients have not contracted Legionnaires' disease also yield L. pneumophila. Aerosolized tap water from respiratory devices also has been a source of nosocomial Legionella pneumoniaY Tap water in respiratory devices has been reported as a mechanism of transmission of Legionella, and there has appeared to be an association between the disease and drinking tap water. 16 However, recent studies have shown the ubiquity of Legionella in water systems of hospitals and hotels.

It has been suggested that in the absence of cases of Legionnaires' disease, major attempts to eradicate the organism should not be undertaken. Clearly, however, when nosocomial legionellosis is prevalent, it is necessary to eliminate the organism from the environment. The Iowa study is a classic example: identification of Legionella in the water system of patients in the hematology-oncology unit resulted in restriction of water for these patients and not allowing the patients to shower.

There are a variety of disinfectants available for cooling tower water contaminated by Legionella but these are not useful for potable water. 2s Ozone will not eradicate L. reported. 3°a This resulted from contamination of the water bath in which the patient received daily physical therapy. The organism survived even though povidone-iodine was used to disinfect the tank. Povidone-iodine at concentrations of less than 1,000 ppm will not eliminate Legionella.

There is no evidence of person-to-person spread of Legionella, and guidelines from the CDC recommend taking precautions with secretions from patients with legionellosis. 31

The key to the prevention ofLegioneUa nosocomial infection is the recognition of disease caused by the microorganism. In the past, this was extremely difficult since diagnosis was characteristically based on serologic titers, which were not available until 3 weeks after the infection developed. It is now feasible to grow Legionella in media available in most microbiology laboratories, with isolation of the organism within several days of submission of a specimen. Thus, development of an unexplained pneumonia in a patient with a hematologic malignancy, and particularly in patients with renal transplants, cardiac transplants, or bone marrow transplants, from whom no microorganisms are isolated, should prompt an attempt to isolate Legionella. Once sputum or bronchial samples are obtained for culture for Legionella, it seems reasonable to begin a chemotherapeutic trial of erythromycin, since this antibiotic is infrequently utilized as initial therapy in these kinds of patients but has proved to be quite effective in the treatment of Legionella infection. Once Legionella is diagnosed, the patient's environment should be carefully evaluated to prevent further episodes occurring within the hospital. The development of several cases should cause one to ensure that patients at high risk avoid any aerosolized water while the potable water and shower systems and air conditioning are being evaluated. Windows in hematology-oncology units and renal transplant units should not be opened to air from areas in which there is water or construction.

Mycobacterium fortuitum and Mycobacterium chelonei are rapidly growing mycobacteria that appear in culture in 3-14 days. These organisms grow on a variety of different agar media and are nonpigmented. They can be distinguished from each other because M. fortuitum will reduce nitrate and take up iron. Both organisms are arylsulfatasereaction positive. Nosocomial infections caused by these organisms have been known since the 1960s. 32 Occasionally nosocomial infection has followed corneal surgery or abscesses have formed at the site of injections. Wound infections following vascular and cardiac surgery have developed at the sites of aortic grafts. Outbreaks of M. chelonei wound infection have followed varicose vein stripping operations. Hard nodules with surrounding areas of inflammation develop 3 weeks to 3 months after the operation. The lesions become fluctuant and drain pus, which ultimately will yield mycobacteria. These outbreaks have been traced to contaminated skin disinfectant used in presurgical care.

There have also been sporadic cases of sternal wound osteomyelitis, pericarditis, and even mediastinitis following cardiopulmonary bypass surgery, aa These outbreaks have been caused by both M. chelonei and M. fortuitum. 34 ' 35 The illness characteristically manifests 3 weeks after surgery with moderate tenderness and erythema along the sternal incision site. Material drained from the area characteristically has been nonpurulent, and the instability of the sternum has suggested the presence of sternal osteomyelitis. Cultures frequently have yielded normal skin flora, and the wounds characteristically have failed to respond to extensive debridement. In spite of extensive investigation, the sources of these various outbreaks have never been defined. It has been postulated that bone wax used at the time of surgery may have been the source of contamination. Sternal osteomyelitis has also occurred after open heart surgery without valve replacement and has been associated with the use of bone wax. 35 Four of 19 patients in one outbreak died of uncontrolled infection. M. fortuitum has also caused infection at Hickman catheter sites in immunocompromised patients. 35a

M. chelonei has contaminated porcine heart valves that have been preserved with glutaraldehyde. 36 In spite of a significant degree of contamination of valves with these organisms, the majority of patients who received the valves were unaffected, although pericarditis and abscesses did develop in some of the patients. The risk of contamination of porcine valves has been significantly lowered by improved techniques of processing the valves; nonetheless, culture at the time of valve implantation is still recommended. Such cultures must be incubated at 28 ° C and held for at least 4 weeks. M. chelonei endocarditis has been demonstrated up to 4 months after heart valve placement with metallic valves, indicating that airborne contamination probably also occurs on occasion.

Augmentation mammoplasty has also been complicated by M. fortuitum and M. chelonei infection. The signs of infection may appear within 1 week but more typically appear much later, sometimes as late as a year after implantation of silicone gel prostheses. Typically, the augmented breast becomes painful and swollen. There is no erythema or drainage. Patients do not have fever or signs of systemic infection. However, on incision of the breast, serosanguinous or purulent material is found in the pocket surrounding the prosthesis. The material has no odor, and Gram stain characteristically will show many white blood cells but few organisms.

Disseminated M. chelonei infection has followed renal transplantation. 37 The patient usually presents with multiple subcutaneous nodules that are tender, red, and elevated, and situated primarily on the lower extremities. Osteomyelitis has also developed in this situation. M. chelonei also has been a cause of peritonitis in patients on long-term peritoneal dialysis. 3s In these situations, infection from this organism cannot be differentiated from that caused by other microorganisms. Typically there is abdominal pain, low-grade fever, and cloudy peritoneal fluid. Contamination of dialysis machines may lead to multiple infections because of cross-infection in a dialysis center.

The important consideration with these atypical mycobacterial organisms is to realize that although they are rapidly growing organisms in terms of mycobacterial growth, they grow slowly compared to ordinary bacteria. Furthermore, these organisms do not stain well with standard Gram stain procedures, and they may be confused with skin diphtheroids. Wound infections at the site of injection, cardiac sternotomy infections, and dialysis infections that fail to yield bacteria should lead one to suspect the presence of M. fortuitum or M. chelonei. Acid-fast stain procedures should be carried out and the laboratory notified that culture material should be kept for prolonged periods. These organisms occasionally take 2 weeks to grow on blood agar and an equally long time on standard media for mycobacteria. The initial appearance of the organism on agar plates may suggest diphtheroid contaminants, which the laboratory would discard. If the laboratory has been notified of the suspicion of Mycobacterium, proper acid-fast staining procedures will be carried out.

Therapy for these infections is exceedingly complicated and usually unsatisfactory, especially in immunocompromised patients. Although wound infections in a normal host usually heals following drainage and debridement, it may be necessary to treat with antimicrobial agents. M.

fortuitum and M. chelonei are resistant to isoniazid, rifampin, ethambutol, and streptomycin. Some strains are susceptible to amikacin and doxycycline. 39 Some strains are also susceptible to trimethoprim-sulfamethoxazole, erythromycin, or cefoxitin. Unfortunately, there are no wellstandardized tests to determine the susceptibility of these organisms. The new Bactec radiometric susceptibility test system may provide susceptibility information more rapidly so that it can be used in planning a course of treatment. 4° The appearance of mycobacterial infections in several hospitalized patients should prompt an immediate investigation to determine a possible source of contamination in supplies used in the treatment or surgical care of these patients.

Recently xenopi antigens. Patients were successfully treated with isoniazid and streptomycin since the organisms were uniformly susceptible to readily achievable concentrations of these agents.

The availability of new culture techniques for Mycobacterium should make it possible to recognize more readily mycobacterial infection.

Nosocomial infection caused by M. tuberculosis has occurred. The greatest hazard is to newborn infants and immunocompromised patients who are exposed to a health care worker with active disease. Tuberculosis is infrequently acquired from other patients. Burk et al. 42 reported that 514 infants exposed to a nursery supervisor with active pulmonary tuberculosis and cough did not develop the disease. They postulated that the ultraviolet light and frequent air changes prevented the infants from acquiring the disease, but it is equally likely that as a supervisor, the nurse had less direct contact with the infants than an ordinary nurse would have. difficile, and it has also been isolated from the urethra or vagina of patients attending venereal disease clinics. Fekety and colleagues, 44 using selective media, found that C.

difficile is common in the environment of patients in the hospital who have diarrhea caused by the organism. It has been found on the floors, toilets, bedding, mops, furniture, scales, and hoppers. However, the air, food, and walls in patient rooms have been negative. C. difficile has been isolated from the hands and stools of asymptomatic hospital personnel.

Since antibiotic-associated colitis has occurred in outbreaks in hospitals in Dallas, St. Louis, Chicago, and Birmingham, England, it is suggested that there is nosocomial spread of the organism within the hospital environment. The epidemiology of the spread is not clearly established, but patients, hospital personnel, and the inanimate environment are probably the most important sources of the organism. Patients receiving antibiotics acquire C. difficile from the hands of personnel who attend other patients in the environment. The number of clostridia needed to infect human subjects is not known, but only a few organisms are needed to colonize antibiotic-treated hamsters. Patients with diarrhea shed significantly more organisms than do asymptomatic carriers. Thus, patients with antibiotic-associated diarrheal disease should be isolated and enteric precautions followed to prevent infection. A recent report by Savage and Alford 45 indicated that two patients who shared a commode chair during hospitalization developed antibiotic-associated pseudomembranous colitis. It is extremely important that potentially contaminated objects be properly cleaned, and that careful hand-washing techniques be stressed in areas in which patients with this illness are present.

Investigation of outbreaks of diarrhea caused by C. difficile by Tabaquchali et al. 46 demonstrated that in oncology and orthopedic units, the same strain was isolated from the patients and their environment. This provides strong evidence for cross-infection between patients and that the organism is hospital-acquired. It also emphasizes that C. difficile is a nosocomial infectious agent and that isolation precautions should be instituted, particularly in areas where patients are at risk. The finding that certain C. dif-ficile serogroups were associated with outbreaks of pseudomembranous and antibiotic-associated colitis, whereas other groups (a, b, c, and d) were isolated mainly from mothers and newborn infants, suggests that certain strains may have a greater virulence, and may explain differences between symptomless carriage of C. difficile in infants and the severe pseudomembranous colitis seen in children and adults receiving antibiotics. Particular attention must be given to the proper cleaning of proctoscopic and colonoscopic equipment, since this type of equipment may become contaminated and be the source of outbreaks in a hospital.

It is clearly more important to determine whether C. difficile toxin is present in a patient's stool when he or she has unexplained diarrhea while receiving antibiotics in the hospital. C. difficile is much more likely to be the pathogen than Salmonella or ShigeUa.

Treatment for C. difficile diarrhea entails stopping the antibiotics and determining if the diarrhea will cease. If diarrhea does not cease, oral vancomycin in a dosage of 125 mg every 6 hours for 7 days will be effective. Metronidazole, 500 mg PO every 8 hours, is also effective and is much less expensive. Relapses occur in 10%-15% of patients, so enteric precautions should not be discontinued prematurely. them from other corynebacteria. These include catalase positivity, lack of motility, nitrate urease and oxidase negativity, and acid production from glucose but not from sucrose. In most of the studies, the antibiotic-resistant organisms have been recovered most frequently from rectal and groin cultures, although they have also been found in the nose and throat as well. Infection caused by corynebacteria characteristically follows a break in a skin barrier where the organism may cause infection, either singly or with other agents. Characteristic sites of infection have been intravenous (IV) catheter insertion sites, bone marrow biopsy sites, adhesive tape removal sites, or sites of skin wounds of other types. These organisms also have the ability to cause endocarditis. A combination of prolonged granulocytopenia together with a break in mucocutaneous barriers, such as that incurred in the placement of a Hickman or Broviac catheter, sets the stage for infection by these organisms. Most patients who develop these infections have received broadspectrum antibiotics that remove normal flora and thus allow colonization. Stamm and colleagues 49 have demonstrated that strains have been nosocomially acquired and that cross-infection explains the inhospital acquisition of the strains. Studies from the Baltimore cancer research group have also suggested patient-to-patient spread. 51 The only antimicrobial agent that adequately inhibits these organisms is vancomycin. In situations where an indwelling line is in place, removal of the IV line is necessary, in addition to vancomycin therapy. It is unclear at present whether improved bathing techniques will decrease colonization with this organism and subsequent infection.

In contrast to the situation obtaining with other organisms, it does not appear that staff members participate in dissemination of this organism, but rather one patient infects another. However, bed sheets and monitoring equipment in the rooms of patients colonized with these organisms may be contaminated, and the organism may by this means contaminate other susceptible patients within the environmentY' 52 Microbiology laboratories must be alerted to the importance of these organisms when isolated from blood cultures of patients who are granulocytopenic or receiving long-term IV therapy. Although most cultures become positive in 24-72 hours, in some situations it may take as long as a week for the blood cultures to become positive. 52 Prevention of infection caused by antibiotic-resistant Corynebacterium JK will most effectively be achieved by scrupulous attention to IV sites in all granulocytopenic patients, since colonization with this organism may persist for months and infect the venous site if it is not properly cared for.

Serratia marcescens has become a formidable nosocomial pathogen. This gram-negative bacillus is naturally resistant to a number of older antimicrobial agents. It also has the ability to acquire resistance to many antibiotics via plasmid mechanisms, and most recently it has been shown to become resistant to the ~-lactamase-stable cephalosporins. ~s In general, the patients who become infected with Serratia tend to be older, have been hospitalized for longer periods of time, and frequently have been treated with multiple antibiotics. 54 Urinary catheterization is a significant risk factor for the development of urinary tract infections caused by Serratia, and prior surgery may contribute to the recovery of Serratia from the respiratory tract. 55 Serratia has also been found to contaminate burns and surgical wounds. There appears to be frequent cross-infection from a hospital reservoir of resistant organisms in patients who remain in the hospital for long periods of time. ~6 Nosocomial infections of the urinary tract have been associated with urine-measuring containers and urometers. In a study by Rutala et al., 57 it was demonstrated that the acquisition of an epidemic strain of Serratia was associated with care in the intensive care unit, the presence of an indwelling bladder catheter, treatment with antibiotics, and exposure to devices used to measure specific gravity and urine volume. The same organism was recovered from the handwashing fluid of nursing personnel, and it was postu-lated that the urometers and urine-measuring containers served as inanimate reservoirs for the resistant Serratia, which was subsequently inoculated onto the hands of medical personnel or directly into a catheterized patient.

On our neurosurgical service, we have encountered an identical problem related to urine-measuring containers that were not autoclaved following use. In the reported outbreak, as well as in the outbreak at our hospital, disinfection procedures for the urometers and urine-measuring containers eliminated the epidemic, and additional cases of drug-resistant Serratia were not seen after the institution of routine disinfection of the inanimate reservoirs.

During the late 1970s gentamicin-and tobramycin-resistant strains were encountered in many institutions. The emergence of the gentamicin-and tobramycin-resistant Serratia paralleled increases in the use of these aminoglycosides. In general, resistant strains have been found primarily in patients with prolonged use of indwelling urinary catheters, as Amikacin-resistant Serratia organisms have also been encountered in some institutions, although distinctly less often. It is critical to recognize early the appearance of Serratia as a cause of infection in a unit within the hospital, since patients with indwelling catheters and this organism should be isolated. Careful attention to the disposal of urine from these patients and to handwashing by personnel in contact with the patients will prevent further spread of the organism. Such measures are particularly important on urology services or on services such as neurology or neurosurgery, which treat large numbers of patients with spinal cord injury who require indwelling urethral catheters. Since outbreaks have been related to contaminated inanimate objects, a careful investigation of the role of equipment in spreading this organism is essential. S. marcescens has also occurred in neonatal units, where the source was colonized symptom-free babies. Despite adequate handwashing by staff, babies have become contaminated and infection resulted, perhaps from overcrowding. 58a

Providencia stuartii is another organism that has been associated with nosocomial urinary tract infections. This gram-negative organism is a fairly infrequent cause of infection within the hospital. However, like other organisms, it can be transferred from one patient to another. In a study by Whiteley et al., 59 two episodes of P. stuartii infection occurred, involving 30 patients in one unit and 11 patients in another. The strain was transmitted when a patient was moved from one unit to the other. In a study by Penner et al., 6° serotyping also revealed that over a 9month period, three episodes of infection occurred in two adjacent units. Providencia was introduced into the hospital from a patient transferred from another institution. This illustrates the necessity of doing cultures on urine samples obtained from patients transferred from chronic care facilities where the patients may have received multiple antibiotics, selecting out bacteria resistant to several antibiotic agents. The most serious nosocomial infections which have been caused by Providencia have been those with bacteremia, usually secondary to urinary tract infection. 61 In most instances bacteremia has followed manipulation of an infected urinary tract. The signs and symptoms of Providencia bacteremia are similar to those of other gram-negative bacteria. As with Serratia, nosocomial acquisition ofP. stuartii resistant to several antibiotic agents has been related to length of hospital stay. 62' 63 Bacteria resistant to gentamicin, tobramycin, and carbenicillin have been reported. Fortunately, most Providencia species are susceptible to amikacin and to the new ~-lactamase-stable cephalosporins, such as cefotaxime and ceftizoxime.

Morganella morgani and Proteus mirabilis have been reported to be causes of serious nosocomial septicemia in a cardiac surgery unit. 64 There was cross-infection of patients with serious wound infections, and three patients died of septicemia. P. mirabilis probably is a cause of nosocomial infection more frequently than is realized, but lack of distinctive antimicrobial susceptibilities causes the organism to be overlooked. 65 Citrobacter diversus has been a fairly infrequent cause of nosocomial infection. It has, however, been associated with serious infection, namely, sepsis, meningitis, and brain abscess in the newborn. Parry et al. 6~ reported an outbreak of C. diversus infection in a 350-bed community hospital. Two infants developed sepsis and meningitis and nine additional infants had asymptomatic umbilical colonization. The infants that developed this disease did not differ from control or noncolonized infants with respect to clinical background or environmental variables. Cultures of nursery personnel identified a hand carrier who had a marked dermatitis from repeated handwashings. This nurse used handcreams and plastic gloves overnight, maintaining the Citrobacter around the sites of rings. Removal of the hand carrier resulted in elimination of neonatal colonization and decreased the number of other enteric bacteria found on umbilical stumps. Transmission within the nursery appeared to have been from the nurse's hands to the infant's umbilicus. Manipulation of umbilical stumps may have been responsible for dissemination of the microorganism. It is interesting that the use of triple dye on umbilical stumps and chlorhexidine handwashing preparations did not eliminate the microorganism.

In another nosocomial epidemic, five infants born over a 2-year period developed meningitis caused by serotype 02 C. diversusF Four infants had brain abscess caused by this organism. Brain abscess, particularly with porencephalic cyst formation, is characteristic of C. diversus infection of the newborn. One colonized infant remained in the hospital for the entire 2-year period and may have been the source of the colonization of other infants. Six nurses were also found to be colonized during the time of the epidemic. Infants who became colonized were distinguished by intensive care therapy, gavage feeding, and perinatal distress. In this epidemic there probably was a fecal reservoir with person-to-person transmission of C. diversus. The cost to control the outbreak was analyzed to be approximately $110,000.

C. diversus is the third or fourth most frequent cause of neonatal meningitis. When the agent appears in a neonatal unit, studies should be undertaken to determine whether long-term patients or personnel are the source of the infection. Unfortunately, chemotherapy for this infection has not been very satisfactory, since by the time the meningitis has been recognized, brain abscesses are frequently present. Optimal therapy for this infection is not known. However, the new third-generation cephalosporins, cefotaxime or ceftizoxime, may prove beneficial since they achieve high concentrations within the CNS and are highly active against C. diversus.

Enterobacter sakazakii has been reported to be an important but uncommon cause of neonatal meningitis. In a report by Muytjens et al., 67a positive cultures were obtained in four patients in one hospital in the Netherlands. It is possible that the infection was transmitted by formula. Like C. diversus meningitis, E. sakazaki meningitis in the newborn has a dreadful prognosis. The organism should be differentiated from E. cloacae and attempts made to determine an environmental source when a case occurs in a nursery.

Pseudomonas maltophilia has been the source both of true nosocomial bacteremia and of pseudoepidemics. 6s' 69 Fisher et al. 6s isolated P. maltophilia from intraoperative blood cultures in 8 of 13 children undergoing open-heart surgery during a 5-week period. The outbreak was traced to contamination of the calibration device used on the pressure monitoring system and the sensor surfaces of transducers used in the system. Although the transducer membrane was intact, reflux of fluid into the monitoring line occurred, resulting in the contamination. Sterilization of the transducers and revision of the calibration device abruptly terminated the outbreak. Other organisms have contaminated pressure dome devices, and the appearance of an unusual organism in patients in whom these devices are used should suggest the possibility of contamination. 69 Conversely, Semel et al. reported a P. maltophilia pseudosepticemia outbreak. During a 17-month period, 25 hospitalized adult patients had blood cultures positive for P. maltophilia. Review of the patient's hospital records showed that these organisms were contaminants and that blood for coagulation studies and for cultures that were subsequently positive had been drawn simultaneously. The source of the contamination was the black-top evacuated collection tubes used for coagulation studies in adults. Inoculation of contaminated black-top tubes prior to the inoculation of blood culture bottles would yield false positive blood cultures, hence, pseudosepticemia. However, one patient undergoing therapy for streptococcal infection of a prosthetic heart valve, who had frequent coagulation studies done, was found at autopsy to have superinfection of the prosthetic heart valve with P. maltophilia. Presumably, infection occurred after reflux of contaminated anticoagulant from the evacuated collection tube into the vein. Thus, contaminated collection tubes are a potential source of confusion in the diagnosis of infection as well as a potential source of true infection.

Pseudomonas aeruginosa is well recognized as a common cause of hospital infections. 7° Sherertz and Sarubbi 71 reported that P. aeruginosa caused 5.3 infections per 1,000 patients at a university teaching hospital in a 3-year study period. Residence on surgery or medicine services, advanced patient age, exposure to burns, and intensive care correlated with a high incidence of infection. The most common sites of P. aeruginosa infection have been the lower respiratory tract, urinary tract, bloodstream, and surgical wounds. Nosocomial Pseudomonas infection of the lower respiratory tract and bacteremia frequently follow stays on intensive care units, whereas urinary tract infections caused by this organism are seen on neurosurgical and neurology services. Drug resistance to both the antipseudomonas penicillin and to aminoglycosides is common.

Hilton et al. 72 reported a cluster of ten nosocomial eye infections in three intensive care units during an 18-month period. Nine of the ten patients were intubated and all were obtunded and had copious sputum production. P. aeruginosa was involved in six of the infections, including the most severe, which involved corneal ulcers and corneal rupture. Study of the bacterial dispersion that occurred during tracheal suction showed that patients with copious secretions had high numbers of bacteria, which could infect a patient's eye as nurses withdrew catheters diagonally across the patient's face. We have recently seen six patients who developed Pseudomonas septicemia in an intensive care unit; the infection was related to contamination of necklines. The patients had large volumes of secretions at a tracheostomy site or nasotracheal tube site which contaminated an indwelling line in the neck.

Prolonged outbreaks of nosocomial urinary tract infection caused by Pseudomonas have been associated with contaminated urine-measuring containers and urometers used in surgical and intensive care units. 73 Urinary drainage bags and external urinary sheet catheters also are frequently contaminated with Pseudomonas. which results in spread of the organism to other patients. 74 As illustrated by these examples of nosocomial infection caused by Pseudomonas, close attention must be paid to the care of patients in intensive care units and to the inanimate objects that come in contact with patients receiving antibiotics that will eliminate all bacteria except Pseudomonas. There are multiple sources of Pseudomonas in the hospital (Table 3) .

Nosocomial diarrhea caused by Salmonella is a relatively uncommon event. Severe disease was seen in the late 1960s from contamination of eggs with Salmonella. The eggs were either inproperly cooked or used raw and caused serious outbreaks, primarily in hematology-oncology patients. A new nosocomial hazard has been the spread of 75 76 Salmonella infections by fiberoptic endoscopy. ' These pathogen associated with bacterial rot of onion bulbs. During the 1960s and early 1970s it was recognized that this organism could use a wide range of organic compounds as energy sources for growth. It is now known that several organisms, including P. kingii and P. multivorans, are the same as P. cepacia. P. cepacia is found in many soils and waters but is found increasingly commonly in hospitals. The only two common community-acquired infections caused by P. cepacia are endocarditis occurring in drug addicts and dermatitis occurring in troups who have had their feet immersed in contaminated waters for long periods of time. In the hospital, P. cepacia colonizes wet surfaces in aqueous solution. Since its invasiveness is fairly low, the patients infected are primarily those who require instrumentation of one sort or another, or those who have serious underlying diseases. The intact skin and its normal bacterial flora tend to pro-tect quite well against P. cepacia. P. cepacia has been implicated in a number of hospital epidemics and in isolated cases of nosocomial infection, sl-s7 Epidemic bacteremia has been traced to contaminated normal saline, human serum albumin, injectable anesthetics, detergents used to clean pressure transducers, aqueous benzalkonium chloride solutions, aqueous chlorhexidine solutions, and contaminated water baths used to thaw frozen blood products or to warm them. Epidemic infection of wounds, the respiratory tract, and the urinary tract has resulted from contamination of aqueous chlorhexidine, benzalkonium chloride, topical tetracine, disposable catheter nebulizers, and mist therapy units, as well as cocaine and saline solutions.

In the majority of nosocomial P. cepacia infections the common link is a contaminated water supply, which is related to the ability of this organism to grow and multiply in distilled water or disinfectants as well as it can in the broth media used in the laboratory. Furthermore, P. cepacia is extremely resistant to the majority of antimicrobial agents. It is not inhibited by most aminoglycosides or by the antipseudomonas penicillins. The organism also produces a cephalosporinase that allows it to destroy a number of the new third-generation cephalosporins. Indeed, the organism is able to grow with penicillin as its sole carbon source. The ability to survive and multiply in water, a nutritional versatility, and resistance to antibiotics and the majority of traditional disinfectants make this a formidable organism when it contaminates the hospital environment.

P. cepacia probably colonizes many more individuals than it actually infects. This is a significant problem, since routine surveillance of nosocomial infections may not identify P. capacia in the environment and a major epidemic can occur very rapidly. In addition, a number of "pseudoepidemics" have been attributed to this organism. These usually occur through contamination of devices used to draw blood or samples for culture, material which is then sent to the laboratory. This may result in inappropriate use of antibiotics since the physician will believe that the patient is infected with this microorganism. In evaluating P. cepacia infection in the hospital, close attention must be given to all water or fluid sources that may have been con-taminated by the organism. Successful control measures include sterilizing contaminated solutions and eliminating contaminated apparatus that may be used in catheterization or other therapeutic procedures. Fortunately, most P. cepacia organisms are inhibited by trimethoprim-sulfamethoxazole.

Staphylococcus aureus resistant to methicillin and to the other penicillinase-resistant penicillins was recognized in Europe in the 1960s. Outbreaks of nosocomial infection occurred in many European hospitals, and staphylococci resistant to several antibiotic agents persisted as endemic nosocomial pathogens. Infection in the hospital setting caused by methicillin-resistant S. aureus was distinctly uncommon in the United States. Although methicillin-resistant staphylococci are called penicillin resistant, in fact these organisms are resistant to all ~-lactam antibiotics, including the most recently released third-generation cephalosporins, and even to the thiemamycin compound, imipenem, and to the penems that are under investigational study.

Methicillin-resistant, or more properly ~-lactam-resistant, S. aureus infections occur primarily in large tertiary. referral hospitals affiliated with medical schools. 89-~ Smaller hospitals in the community rarely encounter these organisms. Why methicillin-resistant S. aureus is confined to large medical school-affiliated hospitals is not clear. However, these institutions tend to have patients at high risk for the development of infections caused by this organism, namely, patients in special care areas such as burn, intensive care, and trauma units. Introduction of the organism directly from the community is distinctly uncommon with the exception of the drug addict. Methicillin-resistant S. aureus appears to be introduced into the hospital by patients who are infected or colonized at other hospitals that have already experienced methicillin-resistant S. aureus outbreaks. Other sources of infection are narcotic addicts, who frequently are infected or colonized by these organisms. Hospital personnel are involved in the dissemination of methicillin-resistant S. aureus, and transient carriage on the hands of personnel appears to be the most important mechanism of transmission from infected and colonized patients, who remain the institutional reservoir. s9

Once introduced into the hospital, methicillin-resistant strains of S. aureus often become established as endemic nosocomial pathogens. In some institutions the overall incidence of staphylococcal nosocomial infections does not increase when methicillin-resistant staphylococci are introduced, s9 but in others there is a great increase, as Bocce et al. demonstrated for the University of Mississippi. 91 Patients with burn wounds are at particularly high risk for acquiring methicillin-resistant S. aureus infection, and several outbreaks have been associated with burn wound contamination. Burn wound patients also may be a source of contamination of the inanimate environment, since organisms have been cultured from hydrotherapy facilities and from operating rooms in which burn patients colonized or infected with methicillin-resistant S. aureus have been treated. 92 Fortunately, in other patients care areas, such as surgical and intensive care units, inanimate contamination is uncommon. 93 The role of nasal carriage during hospital outbreaks of methicillin-resistant S. aureus is not clear. In most institutions, even during periods of outbreak caused by these organisms, less than 1% of employees carry methicillin-resistant strains in their nares, s9 However, postoperative infections have been associated with nasal carriage by nurses or surgeons, and nasal carriage by hospital personnel may influence the epidemiology of some hospital outbreaks.

The severity of staphylococcal infections caused by this organism are illustrated by an outbreak that occurred at the Harborview Medical Center, Seattle. A patient was transferred from a burn unit in another state to the hospital. Despite the standard wound precautions, a methicillinresistant S. aureus was transmitted to 34 patients in the subsequent 15 months. Twenty-seven patients were infected with diseases such as pneumonia, empyema, bacteremia, endocarditis, osteomyelitis, and burn and wound infections. Seventeen of the 34 patients died. Most patients who became colonized and infected were debilitated by burns, major trauma, or surgery that required prolonged hospitalization in intensive care units, and they frequently had received other antibiotics. This epidemic, like several other recent epidemics, illustrates that the methicillin-resistant staphylococci now being encountered are as virulent, as was true in the past.

The most important way to control infection caused by methicillin-resistant S. aureus is to isolate patients known to be colonized or infected by the microorganism. Patients at high risk for acquisition of methicillin-resistant S. aureus are patients with cutaneous wounds and those who are receiving antibiotics. Patients with colonized or infected wounds (in whom direct-contact transmission is the most reasonable mechanism of spread) should be managed with wound and skin precautions. Patients with extensive burn wounds or lower respiratory tract infections require strict isolation since there is the potential for airborne transmission. Patients who are colonized or infected on mucosal surfaces or in the urinary tract can be managed with strict handwashing precautions after direct contact. Precautions must be maintained for the duration of the hospitalization. 9a Wound colonization often persists after antibiotic treatment for staphylococcal infection and may last for several months after hospital discharge. Thus, patients who are known to harbor methicillin-resistant staphylococci should have that fact entered on their charts so that they will be recognized when readmitted to the hospital. Usually attempts to eradicate methicillin-resistant S. aureus from colonized patients have not been successful. The best management of employees who are found to carry methicillin-resistant S. aureus is unknown. An employee carrier should be removed from direct patient care activities if he or she has been implicated in nosocomial transmission. A combination of oral rifampin and trimethoprim-sulfamethoxazole may prevent the emergence of rifampin-resistant mutants and allow nasal carriage eradication.

Handwashing is particularly important in preventing the spread of methicillin-resistant S. aureus because of the organism's propensity to colonize and infect wounds and because the most likely mechanism of transmission is transient carriage on the hands of hospital personnel. Experi-ments have demonstrated that even brief handwashing after contact with methicillin-resistant S. aureus wounds will prevent spread of the organism. Finally, it is extremely important for the microbiology laboratory to perform antibiotic-susceptibility tests in a manner that will detect these strains. If methicillin-resistant staphylococci are to be detected, organisms should be tested for susceptibility at 30 ° C on agar that contains increased concentrations of NaC1.

At present, the accepted therapy for methicillin-resistant staphylococci is vancomycin. Some success with trimethoprim-sulfamethoxazole has also been noted. Fortunately, a large number of new antimicrobial agents such as the fluorinated carboxyquinolones and coumermycin are active against methicillin-resistant staphylococci and may prove useful in the treatment of infections caused by these important new organisms.

Staphylococcus epidermidis is a ubiquitous skin commensal organism. Thus, finding it often is discounted as contamination or innocuous. Humans are the natural reservoir for S. epidermidis. It is shed from cutaneous sites and contaminates the air, other persons, and environmental surfaces. It is able to remain viable for extended periods because of its resistance to drying and temperature changes. S. epidermidis infections result from contamination of a surgical site by organisms from either the patient's own skin or nasal pharynx or from exogenous sources such as hospital personnel. S. epidermidis caused 8.9% of primary nosocomial bacteremias. 95 It is a significant nosocomial pathogen of wounds and the urinary tract.

The incidence of S. epidermidis bacteremia associated with IV catheter use has increased dramatically in the last few years, particularly with the increased use of Hickman and Broviac catheters and routine subclavian catheterization.

S. epidermidis has accounted for approximately 20% of documented bacteremias associated with these intravascular devices. ~ The most significant problem associated with S. epidermidis nosocomial infection has been infection of pros-thetic devices, such as prosthetic cardiac valves, artificial hips and joints, and cerebrospinal shunts.

S. epidermidis is also a frequent cause of infection of vascular grafts and pacemakers, and it produces peritonitis in patients on peritoneal dialysis. 97 Christensen and colleagues 9s reported 13 episodes of S. epidermidis sepsis occurring over 20 months in 11 patients in surgical and medical care units. The episodes were characterized by fever, toxicity, many positive blood cultures, and colonization of intravascular catheters. There were four deaths, and three patients had multiple pulmonary abscesses. In this particular outbreak, the organisms were characterized by broad-spectrum antibiotic resistance. Study of the epidemiology of the outbreak showed that there was a significant association among strains resistant to several antibiotic agents, prolonged hospitalization, and parenteral hyperalimentation. Most of the patients had been hospitalized in the intensive care unit, and nose and hand cultures of personnel in the unit frequently showed carriage of muitiply resistant S. epidermidis.

Breaks in host defense caused by surgery, catheter placement, prosthesis insertion, or immunosuppression are usually a prerequisite for infection by this organism. S. epidermidis is uniquely able to adhere to plastic and metal surfaces. 99 Organisms isolated from infected shunts or catheter-related sepsis produce an extracellular polysaccharide material that enhances adherence to catheters in vitro and makes the colonies resistant to antiseptics and biocides. Prophylactic and therapeutic antibiotic programs may also predispose to colonization with S. epidermidis and lead to an increased incidence of infection.

The optimal therapy for S. epidermidis infections depends to a great extent on the site of the infection. Minor wound infections can be treated with a penicillinase-resistant penicillin, provided that the organism is susceptible.

Fully 20%-30% of S. epidermidis strains, however, show methicillin resistance, and it is probable that these isolates, like methicillin-resistant S. aureus, are resistant to all ~-lactam antibiotics. Deep tissue infections, such as S. epidermidis prosthetic valve endocarditis or S. epidermidis hip or vascular graft infection, will not respond to cepha-losporin antimicrobial therapy if the organisms are truly methicillin-resistant. Vancomycin is the antibiotic of choice for treating methicillin-resistant S. epidermidis and for empirical therapy of S. epidermidis infections until accurate susceptibility data are available. Studies of prosthetic valve endocarditis have indicated that surgery is frequently necessary in order to achieve cure. The same is true of most vascular graft and prosthetic hip infections. The foreign body must be removed to achieve adequate cure. It has been reported that Hickman line infection and peritonitis with indwelling Tenckhoff catheters can be successfully treated with antimicrobial therapy alone. However, in most cases removal of the catheter is necessary since infections frequently recur.

The prevention of nosocomial infections caused by S. epidermidis is based on close adherence to surgical techniques at the time of prosthetic valve, prosthetic joint, and vascular graft surgery. Similarly, strict attention to the maintenance of IV catheters will reduce the risk of line sepsis, as will close attention to peritoneal dialysis catheters. It is generally accepted that the incidence of infection following orthopedic and vascular surgical procedures can be reduced by administering antibiotics at the time of the procedure. Brief use of a cephalosporin antibiotic on the day of the surgery generally does not result in an increase in methicillin-resistant S. epidermidis. However, prolonged use of a cephalosporin or the penicillinase-resistant semisynthetic penicillins as prophylaxis in orthopedic or cardiac procedures will cause proliferation of methicillin-resistant S. epidermidis, which will produce more nosocomial infections.

The organisms discussed in this section bear no major relationship to each other. They have been grouped merely for convenience to illustrate some points about the organisms.

Listeria monocytogenes is an intracellular pathogen that has been the cause of meningitis in the newborn and of serious meningitis and septicemia in immunocompromised patients, particularly those who have recently undergone renal transplantation. Green reported several cases of nosocomially acquired N. meningitidis group B pneumonitis. The pneumonia apparently resulted from aspiration in an elderly man 2 days after admission to a private room on a general medical ward. Pneumonia developed in a patient in an adjacent room 3 days later. Both of the patients were being cared for by respiratory care technicians, and it was suggested that the organism was transmitted by direct contact, probably by the hands of hospital personnel. Cohen, et al. 1°3 reported the nosocomial transmission of group Y N. meningitidis among a number of oncology patients. These two episodes illustrate that N. meningitidis may be a cause of hospital-acquired pneumonitis.

Branhamella catarrhalis, the former Neisseria catarrhalis, is now recognized as an infecting organism rather than just a saprophytic inhabitant of the upper respiratory tract. Although this organism, like some of the nonpatho-genic Neisseria species, occurs in a large number of normal individuals, there are suggestions that in selected situations the organism can spread from person to person in a hospital setting, producing respiratory illness and otitis in-fectionsJ °4

Ornithosis as a cause of nosocomial infection was recently reported by Broholm et al. 1°6 A patient with a contact history with birds developed a serious respiratory illness and was admitted to a general hospital, where he died. Subsequently 11 persons contracted the disease after contact with the index patient. Eight of the individuals were infectious diseases clinic personnel and one was a patient hospitalized in the same room as the index patient. All patients developed a typical pneumonia and one individual developed encephalitis. Patients were successfully treated with doxycycline and there was no serious sequelae. It was necessary, however, to treat prophylactically 200 healthy contacts, none of whom displayed any symptoms of the disease. This outbreak illustrates the risk to hospital personnel from contact with patients with respiratory disease, and the risk to individuals hospitalized in the same room as the patient. It has been our policy to place in single rooms patients in whom we suspect a diagnosis of ornithosis on the basis of contact with birds and the clinical and radiographic findings.

Nosocomial spread ofHemophilus influenzae type b infection would have been extremely difficult to document in previous years. However, contemporary methods of outer membrane protein subtype analysis have demonstrated the communicability of infection between children in an acute care hospital, l°6 In the past, transmission of infection among children would have gone unrecognized, particularly because the isolates in this particular outbreak differed in susceptibility to ampicillin. However, the outer membrane protein profile of the two isolates was identical and had been observed in less than 2% of type B Hemophilus isolates from patients with invasive disease in that particular area of the country. This outbreak illustrates the ability to use new genetic tools in making an analysis of nosocomial spread of infection, and also illustrates the hazards of relying purely on comparable susceptibility to antibiotics as indicating that isolates are the same or different.

Achromobacter has been reported as an organism contaminating diagnostic tracer material and causing bacteremia. 1°7 Achromobacter has also been reported to have caused an outbreak in an intensive care unit where it contaminated the deionized water of a hemodialysis unit. Thirty-seven patients were colonized, and two fatal infections followed. The organism is inhibited only by ceftazidime and moxalactam. 1°7a Another uncommon organism causing infection has been group E Streptococcus, which has caused obstetrical infections due to contamination of intrauterine pressure transducers, l°s Group A streptococcal infections were common in the past but are rare today. Wiesenthal 1°9 reported a maternal-neonatal outbreak caused by M-untypable T-11 group A Streptococcus which may have been the result of using a single sitz bath.

Streptococcal species other than group A or group B can also cause nosocomial infection. Goldman and Breton reported a series of surgical wounds infections transmitted by an anal, rectal, and nasal carrier of group C streptococci. The surgeon involved had perianal dermatitis, from which the streptococcus was isolated, as it was from his nose and rectum. The patients were successfully treated with antibiotics and the surgeon's carrier state was eliminated with topical bacitracin and oral penicillin and vancomycin. Similar episodes of group B and group A streptococcal infections carried anally have resulted in nosocomial infection.

Most interesting has been the development of nosocomial urinary infection due to Pseudomonas pseudomallei. 111

This organism, found in Southeast Asia, causes cavitary pneumonia or a fulminant form of septicemia. Two patients acquired the organisms after being admitted to a hospital in the region where the organism was endemic. The patients were diabetic and underwent urethral catheterization while they were hospitalized. The isolates of P. pseudomallei from the urine specimens of the two patients were identical to isolates from soil samples taken in the vicinity of the hospital. Fortunately, P. pseudomaUei has not been seen as a nosocomial organism in American hospitals. It is a highly resistant organism that responds very poorly to most antimicrobial agents. One of the most unusual nosocomial bacteremia organisms has been the viridans streptococci encountered in bone marrow transplant patients. It has been postulated that the Hickman catheter insertion sites or oral ulcerations could be the entry points for infection; conversely, in some patients the marrow may be contaminated during manipulation. Nonetheless, it is necessary to be alert to the risk in this special immunocompromised group. (Table 4 ). Aspergillus spores are found universally in unfiltered air because 113 Aspergillus also may be found within the hospital because of the cellulose-based fireproofing material that is used to 114 115 coat steel girders and cement work. ' We have found The most common nosocomial infection caused by Aspergillus is pneumonia, which results from direct inhalation of spores and the subsequent deposition of the spores deep within the lung. Whether there is an intermediate step of nasal pharyngeal colonization is not clear. Spores might be able to proliferate on nasal pharyngeal epithelial cells, particularly in immunosuppressed patients and in patients receiving broad-spectrum antibacterial therapy, which reduces numbers of normal nasal pharyngeal streptococcal flora. Subsequently, mycelial elements from nasal pharyn-~xs geal colonization would be aspirated. Aisner et al.

suggested this might be one mechanism of infection, since prospective nose cultures yielding Aspergillus were highly predictive for the subsequent development of invasive pulmonary aspergillosis. Other studies, however, have not indicated that nasal colonization is an important factor in the subsequent development of either respiratory or generalized Aspergillus disease. Clearly, however, ambient airborne Aspergillus spore content in the hospital is an important factor in the development of Aspergillus infection in the patients. There are reports of a reduction in the incidence of aspergillosis in lymphoma-leukemia patients coincident with moving from naturally ventilated older buildings to buildings with superior ventilation systems. Nosocomial aspergillosis also was noted to have decreased in a Veterans Administration hospital when the new hospital was equipped with an airhandling system that consisted of a roll filter and an electrostatic filter that reduced Aspergillus counts. 112, 119

In our institution, Aspergillus infections declined significantly when all construction sites were monitored for the level of Aspergillus spores and proper barrier systems were put in place to prevent dissemination of Aspergillus into patient areas. A number of techniques have been tried to reduce Aspergillus in the environment. In the University of Maryland Cancer Research Center outbreak, copper-8quinolinolate was used to inhibit the growth of Aspergillus, which had been found in fireproofing material. Although this appeared initially to be effective, subsequently Aspergillus infections again occurred and the organisms could be cultured, indicating that even the use of copper-8-quinolinolate is not totally satisfactory, although it is useful as an industrial fungicide.

Most Aspergillus infection is pulmonary. The usual infection mimics bacterial pneumonia. The typical patient is neutropenic, febrile, often has been hospitalized for a long time, and is receiving broad-spectrum antibiotics. Most patients are receiving some form of immunosuppressive medication. The patient characteristically has fever and dyspnea; cough may or may not be present. Cough, when present, is rarely productive. An initial chest x-ray film may be normal despite complaints of pleuritic chest pain. Patients whose white blood cell count returns to normal have had the best response in this illness. There is no established role for granulocyte transfusions for aspergillosis. If a localized mycetoma is present, it may be necessary to use surgery to remove the lesion, since it can reactivate at a later time.

Cutaneous infections caused by Aspergillus can develop in individuals with underlying white blood cell defects or in patients with a hemotologic malignancy who are neutropenic and receiving broad-spectrum antimicrobial agents. 12° An outbreak of this type occurred in our own institution when seven children developed cutaneous Aspergillus infections at the site of arm board placement or IV line insertion. Arm boards in these children had been in place for 4-14 days. The lesions initially appear as palmar erythema with induration and tenderness, but without fluctuance. Gram stains and cultures of the lesions are negative, although a few polymorphonuclear leukocytes are seen. AspergiUus eventually grows from the lesion. More careful evaluation of necrotic nodules (biopsy specimen taken from the edge of the lesion) and proper staining with KOH and methylene blue will demonstrate the plump, septare, branching hyphae diagnostic of Aspergillus. Fixed stains will show hyphae within the tissue. Since Aspergillus may take some time to grow, it is important to perform adequate microbiologic examinations early in situations such as these where cutaneous Aspergillus is a distinct possibility. In this particular outbreak, the source of the infection was contaminated arm boards and a tape in a storage room in which a leak had occurred in the ceiling, providing a high level of contamination with A. fumigatus, A. flavus, and A. niger. It has been our experience that amphotericin B should be given to patients with cutaneous lesions who are neutropenic, since patients who did not receive amphothericin ~ therapy developed disseminated disease.

Nosocomial Aspergillus endocarditis was seen with greater frequency in the late 1960s and early 1970s. We have seen only one case in recent years, in a pediatric patient who probably developed infection because the ventilation system to the cardiac operating room failed. Close attention must be given to the air intake for cardiac operating rooms, since the long time required for cardiac operations permits contamination of the operative field. Aspergillus-induced endocarditis may not manifest clinically until several months after the operation. Clinical presentation may be initially that of septic emboli to the brain or to a large artery in an extremity because of the growth of a large mycelial vegetation on the heart valve. Unfortunately blood cultures are rarely positive in this illness. The only successful therapy for Aspergillus endocarditis is surgical.

It is important to maintain hospital air as free of Aspergillus spores as possible. HEPA filters remove the majority of fungal spores and should be utilized in any new construction. In areas devoted to immunosuppressed patients, HEPA-filtered air should be utilized. Infection control groups should be aware of all new construction or repairs that will be done in areas in which immunocompromised patients are housed. Cleaning, repair, and maintainance procedures should be organized to minimize disturbance of dust, since vacuuming, air-conditioning maintainance, and dry mopping will increase the number of Aspergiltus spores in the air. Fiberboard ceiling tiles contaminated by water leakage should be replaced, since Aspergillus will grow on these tile areas. Although environmental monitoring in general has not been helpful in reducing nosocomial infections, in those hospitals with large populations of immunocompromised neutropenic patients and in which there is also extensive reconstruction, it will be useful to monitor the spore counts in areas adjacent to reconstruction to prevent outbreaks of Aspergillus infection. Elastoplast wound dressings used following orthopedic surgery. 123 Rhinocerebral zygomycosis has even occurred as a nosocomial problem. 124 Bottone and colleagues 125 reported an outbreak of mucormycosis caused by R. rhizopodiformis at Mount Sinai hospital in New York. Six cases occurred within 9 months. The organism was recovered from adhesive bandages used in a cardiac intensive care unit where a patient developed subcutaneous infection after cardiac surgery. The markedly invasive potential of this Rhizopus strain was manifested by extensive subcutaneous and systemic infection in all six patients, three of whom developed antibodies against the microorganism. Antifungal therapy and extensive surgical debridement were necessary to control the problem.

Another example of a fungal infection acquired from nonsterile dressings has been reported by Boyce et al., 127 who isolated Cunninghamella berthoilletiae from a wound of a 69-year-old diabetic man who developed a gangrenous lesion under a cast. The presence of large nonseptate hyphae in thrombosed blood vessels suggested that the organism played a major role in the development of the infections, and isolation of the organism from the cast padding suggested that wound infection was acquired from nonsterile dressings.

Oberle and Penn 12s reported nosocomial infection caused by Saksenaea vasiformis. This zygomycete fungus is found in soils worldwide. A healthy young man receiving large doses of corticosteroids and antibiotics following serious head trauma developed necrosis of the skin, muscle, tendons and fasciae at an arterial catheter site. The organism was isolated from deep surgical specimens, but it did not produce characteristic sporangia until grown on Czapek-Dox agar. The infectious process resolved without specific antifungal therapy following removal of the arterial catheter. Agger and Maki ~26 reported that lethal nosocomial mucormycosis developed in three previously healthy individuals in an intensive care area while they were being treated for acute hemorraghic pancreatitis, cardiogenic shock, and ruptured abdominal aortic aneurysm, respectively. In two patients the infection was initially noted as a progressive cavitary pneumonia that was refractory to antimicrobial therapy. Mucor was identified in all three patients only at autopsy. Each patient had received large doses of corticosteroids and broad-spectrum antibiotics and had experi-enced respiratory failure, acute renal failure, and severe hyperglycemia in association with total parenteral nutrition. There were no data indicating whether construction was occurring during the period of time of this infection.

These examples of mucormycosis and other fungal infections illustrate the problems that can occur when occlusive dressings are used in patients who, although otherwise healthy, have an organism inoculated on an area of the body in which there has been surgical trauma with disruption of the normal skin barriers. In these situations, initial cultures for bacteria are usually negative, and unless cultures are maintained for an adequate period of time or fungi are specifically sought, the true basis of the infection may be overlooked until a patient dies with systemic disease, alerting the hospital to the presence of an epidemic which hitherto had gone unrecognized.

Nocardia asteroides is an organism widespread in nature. Although Nocardia produces infections in normal individuals, it has become a significant pathogen in the era of renal transplantation and of aggressive chemotherapy for hematologic malignancies and lymphoma. An outbreak of N. asteroides occurred in a nephro-urologic intensive care unit in London in 1979.129 Nocardia was isolated from nephrostomy urine of one patient following several urologic operations; subsequently six patients admitted for renal transplantation over a 3-month period developed proved infections. One patient had an abdominal abscess, whereas the other five had pulmonary lesions. Diagnosis was made in three patients by examination of sputum and in three others by bronchoscopy or examination of pus from an abscess. It is of interest that N. asteroides was found in the dust sampling from the intensive care unit air ducts. The unit was closed, decontaminated with formaldehyde, and subsequently reopened without the appearance of new cases.

Stevens and colleagues 13° suggested that it may be necessary to reevaluate guidelines for isolation of patients with pulmonary nocardiosis. At present, respiratory isolation is not considered necessary. This group reported on seven patients in a renal unit who developed nocardiosis in an interval of 9 months. Six of the patients had received a renal transplant. N. asteroides could be isolated from air and dust within the unit and elsewhere within the hospital. Biochemical and metabolic and immunologic examination of isolates indicated that those from the patients and from the environment were identical and different from the most common type of human Nocardia strains. Both of these reports suggest that, indeed, immunocompromised patients should not be in the same environment as individuals from whom Nocardia can be isolated. On the other hand, there are no data to indicate that routine environmental surveillance for Nocardia would be required in renal transplant units or in oncology units handling patients with Hodgkin's disease or non-Hodgkin's lymphoma. The reports do suggest, however, that following the appearance of a case in a patient who has been in the hospital for a considerable period of time, it would be wise to do some form of environmental sampling. Candida glabrata, formerly called Torulopsis glabrata, has also been recognized as an important cause of fungemia from IV catheters and to produce genitourinary infections on a nosocomial basis. Most of the patients who are infected with C. glabrata have had recent antibiotic therapy and surgery, and many are immunosuppressed. TM This organism has caused endocarditis, endophthalmitis, and disseminated infection. 132 Colonization with the fungus appears to be the primary problem leading to the nosocomial infection when the clinical setting provides the fungus access to body areas from which it is normally excluded. 47 

It is not surprising that only in the last several years has attention been given to viral infections as a major nosocomial disease problem. Difficulties in culturing viruses and performing diagnostic serologic tests were the main reason that viral infections were ignored. In recent years our increased awareness of the spectrum of viral disease, coupled with our ability to culture viruses and the potential for antiviral chemotherapy, has focused attention on nosocomial viral illness. Outbreaks of respiratory and gastroenterologic infection have been most readily diagnosed and are the most clearly defined forms of infection. Most studies have been done in pediatric patients.

With the availability of enzyme-linked immunosorbent assays, rotavirus gastroenteritis has become recognized as a significant hospital problem. For example, Noone and Banatvala 138 reported that 37 patients acquired gastroenteritis caused by a rotavirus at St. Thomas' hospital in London during the period of time that 74 children with acute enteritis were admitted to the pediatric wards. Fifty-nine percent of patients with hospital-acquired diarrheal disease during this time period were ill with rotavirus. Although patients were admitted with a diagnosis of diarrheal disease and stool isolation procedures were used, there was spread of infection. Cubitt and Holzel TM reported an outbreak of rotavirus infection in a long-stay ward of a geriatric hospital. Rotavirus was detected in 47% of symptomatic individuals. In addition, four members of the hospital staff were also infected.

In a study to determine whether human milk would protect against rotavirus infections in children, Berger et al. 135 grouped children entering an infants ward into two groups: those who would receive a normal diet and those who would receive 200 ml of fresh milk a day. In this study there was no difference in infection rate between the two groups, although clinical symptoms seemed to be less severe in children fed human milk. Sixteen of 28 children excreted rotavirus during their stay in one ward, and eight of the 16 developed diarrheal disease.

Rotavirus is clearly one of the most important causes of diarrheal disease in small children under the age of 2 years, When these children are admitted to hospitals, it is extremely important that adequate precautions be taken to avoid contamination of other children or of personnel. Rotation of medical or surgical house staff among different services and rotation of nursing staff may provide a vector for dissemination of rotavirus infections to adults. It is routine to obtain specimens for culture of Salmonella, Shigella, and other enteric pathogens, but more attention must be given to rotavirus infections, in adults as well as children.

Nosocomial viral infections reflect the pattern of the agents in the community. Hence, the viruses that occur in epidemics in the community are the same ones that can be seen on the pediatric ward. t36 The epidemiologic characteristics of nosocomial viral pediatric infections will be quite similar to those of community-acquired infections. Since viral infections of the respiratory tract are so frequent in small children, it is inevitable that a number of children admitted to a hospital may be shedding the virus asymptomatically. It is also important to understand that viral nosocomial infections of the respiratory tract do not necessarily occur more often in immunocompromised children than in normal children. Thus, children on an orthopedic pediatric service or a urologic pediatric service are susceptible to respiratory syncytial virus (RSv), influenza, and parainfluenza viruses, and the other viruses that cause illness in children.

The most important nosocomial infection in very young pediatric patients has been RSv. RSv is the most important cause of lower respiratory tract disease in the first several years of life. The characteristics of the virus that allow it to cause nosocomial infection are (1) the frequent winter or spring outbreaks, (2) the susceptibility of persons in all age groups (with reinfections common throughout life), (3) an immunity of short duration, and (4) the fact that infants admitted with RSv bronchiolitis or pneumonia shed large amounts of the virus for long periods of time, and provide the focus for nosocomial spread of the virus. Despite use of infection control procedures, such as gown-changing, handwashing, and isolation of high-risk patients, nevertheless, in outbreaks of RSv, 26%-38% of patients have developed nosocomial infections with pneumonia, and case-fatality rates have ranged from 7% to 17%. Furthermore, 34%-56% of staff have become infected. Unfortunately, RSv can be communicated both by large droplets or by self-inoculation of contaminated secretions. Hall and colleagues 137 evaluated methods to control the spread of RSv in the infant wards during a community outbreak of infection with the organism. Infants were divided into cohorts and were isolated, with strict attention to handwashing and the use of gowns. Staff who attended the ill infants were also divided into cohorts. Nineteen percent of infants acquired nosocomial RSV disease. Three of 80 developed pneumonia, and one died as a result. Some 56% of 43 staff members became infected, and 82% became symptomatic. Furthermore, four staff members acquired repeated infections after the initial infection. Previous studies from this institution had revealed that 45% of infants, and 42% of staff, in contact with infected infants would acquire nosocomial RSv infections.137, 13s To some extent procedures of isolation will reduce the transmission of RSv to infants, but they do not appear to be very effective with staff, who are often infected by close contact with the infants. Staff members may also inoculate themselves with contaminated secretions.

Studies by Hall and colleagues 1~9 have indicated that inoculation usually occurs through the nose or eye and probably not via the mouth. RSv survives for several hours on gloves and on skin and paper tissue, and up to 7 hours on countertops. Indeed, on the hand, the virus has a shorter survival--one-half hour--than it does on gloves used for iaolation. Studies have indicated that staff who cuddle infected infants or touch contaminated surfaces became infected with RSv. Unfortunately, control of nosocomial spread of RSv has not proved successful, even with division of infants into cohorts and limiting patient contact. However, it is clear that handwashing is extremely important. It seems that the major reason for wearing a mask when handling RSv patients is that it prevents one from touching one's nose with one's hands and therefore spreading the infection to personnel, who would then spread it to other infants 14°. Unfortunately, immunization against RSv is not possible at present. Eriksson and colleagues 141 have shown that RSv may be detected rapidly by immunofluorescence studies. With early detection it may be possible to plan the care of the patient and the personnel in an optimal way to avoid or to maintain nosocomial infection at a low level. It is extremely important that personnel understand that environmental conditions with RSv may permit survival of the virus on their surfaces and skin and that through selfinoculation they will contaminate their nasal surfaces and become infected.

Another use of modern virologic technique is the study of herpes simplex viruses (HSV) by the use of restriction endonucleases, which allows determination of whether cases in a hospital belong to one strain or to a variety of unrelated strains. Buchman and colleagues, 142 studying outbreaks in a pediatric intensive care unit, found there were two independent introductions of HSV-1, resulting in two clusters of epidemiologically related infections. HSV is an important pathogen within the hospital setting, and the use of these techniques may be able to identify individuals who are likely to infect other individuals.

HSV is an important cause of infection in health care personnel, particularly those caring for immunosuppressed patients, who may asymptomatically shed the virus in their oral pharyngeal secretions. Hospital personnel thus may develop a primary infection and may be the source of infection of other patients. HSV has been increasingly recognized as a nosocomial problem. We have recently seen the development of herpes simplex pneumonia following bronchoscopy; the organism probably was introduced deep within the lung at the time of initial bronchoscopy and was found at subsequent bronchoscopy when a marked change in radiographic appearance was noted. This patient responded to IV acylovir therapy with clearing of pulmonary disease and marked improvement in blood gas values. Graham and Snell have outlined the various lower respiratory tract problems encountered with HSV. 143 Varicella is an extremely common nosocomial infection in pediatric institutions. There are many documented reports of spread from a child with either primary varicella infection or from adults or children with varicella zoster.

We have seen numerous cases develop in siblings who were allowed to visit patients, and who then carried the infection to their playgroups. Persistent outbreaks can be maintained for weeks in this fashion. Morens et al. 144 reported an outbreak of varicella zoster infection among patients at the National Cancer Institute. The epidemiologic investigations suggested that the outbreak was caused by two distinct types of disease. One type was acquired without previous exposure to other infected patients and was invariably associated with dermatomal lesions. The other, a typical form, was associated with person-to-person transmission, an equivocal dermatomal distribution, and an incubation period of approximately 11-25 days. The latter probably was varicella occurring in patients who were immunodeficient because of disease, debility, and chemotherapy. This outbreak illustrates that the classic lesions may not appear within the usual time frame of 10-21 days, and that as long as 25 days after an exposure, lesions will develop.

Asano et al. 145 reported an outbreak in Japan in which varicella spread from a child with zoster to three susceptible infants in another room in a children's unit, although they had been strictly isolated. The cases indicate that it is often difficult to predict nosocomial varicella infection or to prevent the spread of the disease simply by isolation in a children's unit. A total of 11 other children on the ward without a history of varicella were given live varicella vaccine before or immediately after the event. None of the children developed symptoms of varicella, and all susceptible children who were vaccinated showed an antibody response.

It is exceedingly important to remember that varicella can be especially severe in adults, in neonates infected in utero, in immunocompromised patients, and in those with preexisting conditions. The recent results of the use of a live attenuated varicella virus vaccine, reported by Weibel et al., 146 indicated that the Oka-Merck varicella vaccine produced few clinical reactions, was well tolerated, caused no virus spread from vaccinated children to sibling controls, and resulted in effective protection against developing varicella. Thus, it appears that we may finally have a solution to this serious problem. The response to vaccination of the immunosuppressed child with leukemia is currently being explored in large cooperative trials. That will be an extremely important factor in preventing the most serious form of this nosocomial infection.

Outbreaks of conjunctivitis caused by adenoviruses, particularly adenovirus type 8, have been well known. More recently, however, conjunctivitis caused by adenovirus type 4 has been reported in hospital personnel who have had contact with a patient with adenovirus 4 pneumonia. 147 The outbreak reported by Levandowski and Rubenis illustrates several important points. The incubation period of the adenovirus 4 virus was 7-10 days, and the illness lasted 5-7 days. The infected virus was regularly present in the affected eyes for 1 week after the onset of symptoms, and visual disturbances persisted in several patients for long periods of time, particularly in one patient, who developed subepithelial deposits. It was interesting that this outbreak occurred after isolation procedures were abandoned in the care of the patient with adenovirus pneumonia.

A patient with a serious adenovirus pulmonary illness will often be admitted to an intensive care unit because of severe problems in oxygen exchange. As a result, other patients in a unit may acquire the disease, or the staff may acquire the disease and communicate it to other patients. More frequently, viral eye infections occur in hospital clinic settings where eye drop solutions have been contaminated by being used in multiple patients.

The question of nosocomial spread of cytomegalovirus (CMV) is an important one because of the risk to nursing personnel who may be pregnant. Primary CMV infection is a major cause of morbidity in renal transplant recipients. Betts and colleagues 14s investigated the risk of transmission of CMV. In a study of 85 patients and 49 personnel, they detected CMV in eight nontransplanted older dialysis patients and 13 patients who had lost their allograft kidneys. Although CMV was present on the unit, no patient or staff member developed primary infection from interpersonal transmission or from transfused frozen red blood cells. All primary infections in renal transplant patients could be accounted for by acquisition from the transplanted kidney. Thus, the dialysis unit and the blood used appear to offer minimal risk to patients and personnel. This will not be the case for newborn babies who are shedding the virus, and the risk from the large numbers of AIDS patients is unknown. Therefore, one should be cautious in handling urine from these individuals until further epidemiologic studies indicate that the risk of transmission via this method is low.

The risk of influenza as a nosocomial infection has long been known. Kapila and colleagues 149 reported an outbreak of influenza from a patient who had been hospitalized for evaluation for other medical problems, but who subsequently developed influenza that had been incubating prior to the hospital admission. Seven compromised hosts in the same unit developed symptoms of pneumonic influenza, and serologic data confirmed influenza A2. Unusual features of the epidemic were the intrahospital localization of the epidemic in compromised hosts, the high rate of pneumonic complication, and the low rate of secondary bacterial infection.

This case illustrates the necessity of paying close attention to the isolation of patients with influenza when they are admitted to the hospital. Rapid spread of influenza makes nosocomial control quite difficult if there is an outbreak. Clearly, control is best achieved by prophylaxis. It is possible to have both chemoprophylaxis with an amantadine and simultaneous vaccination with the current influenza vaccine. Immunization will not be of benefit once a nosocomial infection has been detected because of the very rapid spread of the virus. Hospital intensive care workers, particularly those who work in respiratory care units, should not only be offered vaccination but encouraged to be vaccinated. Unfortunately, it has been our experience that influenza vaccine is infrequently utilized by medical staff when it is offered. A major education program may be required to increase compliance of the hospital staff. In pediatric units, control of nosocomial spread of influenza probably is not feasible since the majority of children will never have had immunization, nor will they have been exposed to viruses of a cross-reacting nature.

Nosocomial spread of parainfluenza viruses, types 1, 3, and 4A, has been well documented. The mode of spread of parainfluenza viruses is not known. 15° It is unclear whether small particles of aerosol in addition to direct contact are important in the spread of the virus. The slow spread of parainfluenza virus during an outbreak does not seem characteristic of the spread of small-particle aerosols, and it is suggested that, like RSV, rhinoviruses, and corona viruses, parainfluenza viruses are picked up on hands and transferred to nasal epithelium, where they produce their infection.

Nosocomial outbreaks of influenza B in the elderly have been reported. TM This virus is spread as influenza A is spread and probably occurs much more frequently than is recognized. Rhinoviruses, which are extremely common during winter months in adult hospital personnel, may be introduced into the hospital, where they will cause nosocomial infection. 152 The infection from rhinovirus in adults usually is mild enough that the individual continues to work or may even be asymptomatic. There is prolonged excretion of the virus, and, like several other viruses, it is spread by self-inoculation and by fomites. 152 Rhinoviruses may be a particularly hazardous virus for the newborn or the elderly in respiratory care units. Control is best achieved by meticulous handwashing. Finally, with respect to enteroviruses, it is important to understand that virus may continue to be shed in fecal material after the patient has recovered from the febrile illness or from aseptic meningitis. Thus, in order to prevent nosocomial spread of enteroviruses in the hospital, patients should be maintained on enteric precautions and strict handwashing techniques are advised for the duration of the hospitalization.

Fortunately, nosocomial parasitic infections are quite uncommon. It is possible for a hospital water supply to become contaminated with Giardia lamblia, but this would be an extremely uncommon occurrence. It is also possible that workers in a food service may shed Entomoeba histolytica and thereby contaminate patients, but again, this is extremely unlikely. One should realize that Strongyloides stercoralis can be communicable if stools with larvae are left at room temperature for 24-48 hours. Thus, it would be possible for a health worker or patient to become infected with this organism. Malaria has been caused by contaminated blood transfusions, but this is an extremely infrequent occurrence at present.

The parasite currently of greatest interest is Pneumocystis carinii. This organism is a major cause of pneumonia in the immunocompromised patient, particularly in patients with autoimmune deficiency syndrome (AIDS). Although it has been established that there is airborne and animal-toanimal transmission of P. carinii, the mode of tradition in humans has not been established. We know that more than two thirds of normal children are seropositive for this organism by age 4 years. It has been the general opinion that most cases of P. carinii infection represent reactivation of latent organisms caused by alteration in the host immunity. Nonetheless, there have been clusters of P. carinii in the United States, which suggests person-to-person transmission. For example, in the 1960s at the Veterans Administration Hospital in Denver, Colorado, there were no renal transplant patients who developed P. carinii infection, whereas ten cases occurred in renal transplant patients at the Colorado General Hospital. 153 In 1975 Singer and colleagues TM at Memorial Hospital in New York reported on three pediatric patients who shared rooms before the onset of infection. The most suggestive cluster of cases was reported by Ruebush et al. 155 from the Riley Hospital in Indianapolis. There was a greater attack rate of P. carinii pneumonia among patients at the Riley Hospital than among similar patients who were part of a multicenter chemotherapy trial at the other hospitals in the city. None of the other patients had been assigned to the same hospital, and on only two occasions had they been admitted to this same unit at the same time. It is also interesting that the hematology-oncology physicians and nurses had a greater seropositivity than phxsicians and nurses in other units. Chusid and Heyrman 156 also reported an outbreak of Pneumocystis pneumonia in a pediatric hospital.

The CDC does not recommend isolation for P. carinii in any situation. However, since many children less than 4 years of age are not seropositive, it seems reasonable not to have patients less than 4 years of age in the same room with a patient with P. carinii pneumonia.

Cryptosporidium is a parasite long known to cause diarrheal disease in animals and fowl. It has recently been recognized as an important pathogen producing serious diarrheal disease in patients with AIDS. It is well known that veterinarians can become infected with Cryptosporidium, and outbreaks of diarrheal disease with this organism have occurred in veterinary schools. One must be extremely careful in handling stool specimens from patients with diarrhea caused by Cryptosporidium. To date, no major nosocomial outbreaks have been recorded, but it seems likely that we will see nosocomial infections with this interesting parasite. Person-to-person transmission of Cryptosporidium has been reported from Great Britain, where a nurse contracted the disease from a child who was infected. 158a

The final unusual nosocomial organism is the mite Sarcoptes scabiei var. hominis, an anthropod of the order Acarina, which produces the highly contagious skin disease, scabies. S. scabiei is an obligate parasite for humans. Scabies affects persons of all ages but it is more common in children in underdeveloped countries. 1~7 It has been a problem in nursing homes, hospitals, and mental institutions. ~58 Gooch et al. 1~9 reported on a nosocomial outbreak of scabies in a 558-bed teaching hospital. There had been reports of scabies throughout Michigan for a year prior to the outbreak. There was an epidemic spread of scabies from a patient to 38 hospital employees and their families and associates. Patients admitted to hospitals from shelters or from nursing facilities in which care is suboptimal may be contaminated with scabies. Unexplained cutaneous lesions compatible with irritation because of infestation with this organism should alert hospital personnel to the problem. There is pruritus, most severe at night. Careful microscopic examination of lesions and microscopic examination of the parasite will identify the problem.

Treatment of scabies includes the administration of lindane, a 1% lotion of cream which is applied overnight to the entire body except the head and followed by washing the next day. All linen and bed garments must be washed in hot water. When a hospital outbreak occurs, it probably is wise to treat all staff who had contact with the infected index patient.

Although great progress has been made in the fight against bacterial infections, and large numbers of effective antimicrobial agents have been discovered or synthesized in the past 50 years, hospital-acquired infectious diseases have remained a significant problem. Major improvements in sanitation and hygiene caused a marked reduction in many community-acquired infections. We have seen the reduction in certain nosocomial infections through strict decontamination of respiratory care equipment, urinary catheters, and vascular catheters. However, some infections in hospitalized patients will not decline as rapidly in the coming decade. The reason is that serious infection, particularly nosocomial infection, in the critically ill patient frequently is different from infection acquired within the community. Infections in the hospital are not just the result of failure in infection control, but of acquisition of bacteria, fungi, or viruses that cause infection or invade the host during procedures carried out to maintain life. The body whose mucocutaneous defenses are bypassed by IV or urethral catheters is much more vulnerable to invasion by pathogens. Therefore, we shall continue to see nosocomial urinary and vascular infections caused by gram-negative organisms and fungi. New bacteria will appear, as noted in this monograph. We know that gram-negative bacteria have surface structures that allow them to adhere to respiratory, gastrointestinal, and uroepithelial cells. The healthy individual is less readily colonized by bacteria or staphylococci. Colonization preceds invasion and infection. The healthy individual may have only a transient bacteremia from an improperly cared-for catheter, whether vascular or urologic, and he or she will be rid of the organism. But patients in intensive care units or individuals with underlying neurologic disease frequently will suffer serious nosocomial infection.

Prevention of viral respiratory disease is extremely dif-ficult in the hospital setting. We shall continue to see outbreaks of RSV, parainfluenza virus, and influenza virus in both pediatric and adult patient care settings. New pathogens such as Legionella and Aspergillus have become significant problems in some institutions. Rapid recognition of the problem and institution of measures to improve water supplies or decrease dissemination of fungal spores will halt such outbreaks. A better understanding of the microbiologic and pathophysiologic aspects of unusual infections acquired in the hospital should aid us in developing more rational approaches to infection. It is hoped that this discussion has provided information to the physician about the less common and more unusual forms of nosocomial disease.

Nosocomial infections in children's wards

Secular trends in nosocomial infections: 1970-1979

Genitourinary infections in the patient at risk: Extrinsic risk factors

Long-term evaluation of decontamination of inhalation therapy equipment and the occurrence of necrotizing pneumonia

Deaths from nosocomial infection: Experience in a university hospital and a community hospital

Hospital-acquired gram-negative rod pneumonias: An overview

Current mechanisms of resistance to antimicrobial agents in microorganisms causing infection in the patient at risk for infection

Comparison of endemic and epidemic nosocomial infections

Acinetobacter calcoaceticus variety anitratus: An increasing nosocomial problem

Nosocomial respiratory tract infection and colonization with Acinetobacter calcoaceticus

A common source outbreak of Acinetobacter pulmonary infections traced to Wright respirometers

Pseudobacteremia: False positive blood cultures from a mist bath contamination

Acinetobacter calcoaceticus outbreak associated with peritoneal dialysis

Acinetobacter calcoaceticus: A nosocomial pathogen with an unusual seasonal distribution

Structure-activity relations of new beta-lactam compounds: In vitro activity against common bacteria

Legionella infections: A review of five years' research

Ecological distribution of Legionella pneumophila

Acathamoebae and environmental spread of Legionella pneumonia

Nosocomial Legionnaires' disease uncovered in a prospective study: Implications for underdiagnosis

Legionnaires' disease associated with a hospital water system: A cluster of 24 nosocomial cases

An outbreak of Legionnaires' disease associated with a contaminated air-conditioning tower

Legionellaceae: Epidemiological link with disease and evaluation of a method for control of nosocomial Legionnaires' disease and Pittsburgh pneumonia

Subtypes of Legionella pneumophila serogroup 1 associated with different attack rates

Legionella pneumonia in guinea pigs exposed to aerosols of concentrated potable water from a hospital with nosocomial Legionnaires' diseases

Legionnaires' disease in a transplant unit: Isolation of the causative agent from shower baths

Isolation of Legionella pneumophila from hospital shower heads

Nosocomial Legionnaires' disease caused by aersolized tap water from respiratory devices

Susceptibilities of algae and Legionella pneumonia to cooling tower biocides

Efficacy of ozone in eradication of Legionella pneumophila from hospital plumbing fixtures

Water-fillings as sources of Legionella pneumonophila in a hospital plumbing system

Legionella pneumophila wound infection

Lack of evidence for person-to-person transmission of Legionnaires' disease

Non-tuberculous mycobacteria and associated diseases

Mycobacterium chelonei iatrogenic infections

Mycobacterium fortuitum epidemics after open-heart surgery

Two outbreaks of sternal wound infections due to organisms of the Mycobacterium fortuitum complex

Mycobacterium fortuitum infection of a Hickman catheter sites

Fastidious mycobacteria grown from porcine prosthetic heart value cultures

Disseminated Mycobacterium chelonei infection: A report of two cases

Peritonitis due to Mycobacterium chelonei-like organism associated with intermittent chronic peritoneal dialysis

Clinical usefulness of amikacin and doxycycline in the treatment of human infection of Mycobacterium fortuitum and Mycobacterium chelonei

Isolation of Mycobacterium chelonei with lysis-centrifugation blood culture technique

Clinical and roentgenographic features of nosocomial pulmonary disease due to Mycobacterium xenopi

Nursery exposure of 528 newborns to a nurse with pulmonary tuberculosis

Antibiotic-associated pseudomembranous colitis

Epidemiology of antibioticassociated colitis

Nosocomial spread of Clostridium difficile

Typing scheme for Clostridium difficile: Its application in clinical and epidemiological studies

Sepsis with a new species of Corynebacterium

Corynebacterium sepsis in oncology patients: Predisposing factors, diagnosis and treatment

Infections due to Corynebacterium species in marrow transplant patients

Characterization and identification of 95 diphtheroid (group K) cultures isolated from clinical specimens

The emergence of Corynebacterium bacteria as a cause of nosocomial infections in compromised hosts

Antibiotic-resistant group JK bacteria in the hospital

SCH 29482 activity against susceptible and beta-lactam resistant mutants of Enterobacteriaceae

Endemic Serratia marcescens in the Veterans Administration Hospital in

The significance of Serratia as an infectious organism

Serratia: Opportunistic pathogens of increasing clinical importance

Serratia marcescens nosocomial infections of the urinary tract associated with urine-measuring containers and urinometers

Epidemiology of gentamicin-resistant, gram-negative bacillary colonization in a spinal cord injury unit

An outbreak of serratia marcescens in a neonatal unit

Nosocomial urinary tract infections caused by two O-serotypes of Providencia in one hospital

Three episodes of nosocomial urinary tract infections caused by one O-serotype of Providencia stuarti

A review ofProvidencia bacteremia in a general hospital

Antibiotic resistance in Providencia stuarti isolated in hospitals

Nosocomial multiply-resistant Providencia stuartii: A long-term outbreak with multiple biotypes and serotypes at one hospital

Serious nosocomial infection caused by MorganeUa morganii and Proteus mirabilis in a cardiac surgery unit

A nosocomial outbreak of infections due to a multiply resistant Proteus mirabilis

Gram-negative sepsis in neonates: A nursery outbreak due to novel carriage of Citrobacter diversus

Epidemic nosocomial meningitis due to Citrobacter diversus in neonates

Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii

Pseudomonas maltophilia bacteremia in children undergoing open heart surgery

Pseudomonas maltophilia

The role of Pseudomonas aerugionosa in infection

A three year study of nosocomial infections associated with Pseudomonas aeruginosa

Nosocomial bacterial eye infections in intensive-care units

Prolonged outbreak of nosocomial urinary tract infection with a single strain of Pseudomonas aeruginosa

Pseudomonas colonization in patients with spinal cord injury

Salmonella typhimurium: Transmission by fiberoptic upper gastrointestinal endoscopy

Outbreak of Salmonella agona infection after upper intestinal fibreoptic endoscpy

Clustering ofAeromonas hydrophila septicemia

Nosocomial Aeromonas hydrophila cellulitis and bacteremia in a nonimmunocompromised patient

Pseudomonas cepacia: Implications and control of epidemic nosocomial colonization

Pseudomonas cepacia as contaminant and infective agent

Pressure transducers as a source of bacteremia after open heart surgery: Report of an outbreak and guidelines for prevention

Contamination of commercially packed urinary catheter kits due to contaminated benzalkonium antiseptic

Pseudomonas cepacia (multivorans) septicaemia in an intensive care unit

Outbreak of Pseudomonas cepacia due to contaminated anesthetics

Pseudomonas species bacteremia caused by contaminated normal serum albumin

Pseudobacteremia caused by povidone-iodine solution contaminated with Pseudomonas cepacia

Pseudomonas cepac/a peritonitis associated with contamination of automatic peritoneal dialysis machines

The emergence of methicillin-resistant Staphylococcus aureus infections in United States hospitals

Epidemiology of nosocomial infections caused by methicillin-resistant Staphylococcus aureus

Multiply antibioticresistant Staphylococcus aureus: Introduction, transmission and evolution of nosocomial infection

Impact of methicillin-resistant Staphylococcus aureus on the incidence of nosocomial staphylococcal infections

Environmental study of a methicillin-resistant Staphylococcus aureus epidemic in a burn unit

An outbreak of methicillin-resistant Staphylococcus aureas in a large medical center

An outbreak of infections 101a. 107a. caused by strains of S. aureus resistant to methicillin and aminoglycosides: II. Epidemiologic studies

Staphylococcus epidermidis infections

Coagulase-negative staphylococcal bacteremia in patients receiving immunosuppressive therapy

Peritonitis during continuous ambulatory peritoneal dialysis

Nosocomial septicemia due to multiply antibiotic-resistant Staphylococcus epidermidis

Adherence and growth of coagulase-negative staphylococci on surfaces of intravenous catheters

Hospital outbreak of Listeria monocytogenes septicemia: A problems of cross infection

Listeria monocytogenes infection in neonates: Investigation of an epidemic

Listeriosis in renal transplant recipients: Report of an outbreak and review of 102 cases

Meningococcal pneumonia: A source of nosocomial infection

Possible nosocomial transmission of group y Neisseria meningitidis among oncology patients

Branhamella catarrhalis as a human pathogen and a possible hospital infectant

Ornithosis as a nosocomial infection

Nosocomial spread of Haemophilus influenzae type b infection documented by outer membrane protein subtype analysis

An outbreak of Achromobacter xylosoxidans sp. related to diagnostic tracer procedures

Nosocomial colonization and infection by Achromobacter xylosoxidans

Nasal colonization of infants with group E streptococcus associated with intrauterine pressure transducer

A maternal-neonatal outbreak of infections due to an unusual group A beta-hemolytic streptococcus

Group C streptococcal surgical wound infections transmitted by an anorectal and nasal carrier

Nosocomial infection due to Pseudomonas pseudomallei: Two cases and an epidemiologic study

Filtering hospital air decreasesAspergiUus counts

Extrinsic risk factors for pneumonia in the patient at high risk for infection

Pulmonary aspergillous during hospital renovation

Aspergillus infections in cancer patients: Association with fireproofing materials in a new hospital

Increased recovery of AspergiUus flavus from respiratory specimens during hospital construction

AspergiUus fumigatus and other thermotolerant fungi generated by hospital building demolition

Invasive aspergillosis in acute leukemia: Correlation with nose cultures and antibiotic use

Decreased frequency of aspergillosis and mucormycosis

An outbreak of aspergillosis in children with acute leukemia

A retrospective review of airborne disease secondary to road construction and contaminated air conditioners

Aspergillosis in 25 renal transplant patients

Hospital-acquired mucormycosis (Rhizopus rhizopodiformis) of skin and subcutaneous tissue

Two cases of rhinocerebral zygomycosis (mucormycosis) with common epidemiologic and environmental features

Rhizopus rhizopodiformis: Emerging etiological agent of mucormycosis

Mucosmycosis: A complication of critical care

CunninghameUa berthoilletiae wound infection of probable nosocomial origin

Nosocomial Saksenaea vasiformis infection

An outbreak of Nocardia asteroides infection in a renal transplant unit

Laboratory evaluation of an outbreak of nocardiosis in immunocompromised patients

Torulopsis glabrata fungemia: A clinical pathogical study

Disseminated Candida glabrata: Report of a uniquely severe infection and a literature review

Hospital acquired rotaviral gastroenteritis in a general paediatric unti

An outbreak of rotavirus infection in a long-stay ward of a geriatric hospital

Effect of feeding human milk on nosocomial rotavirus infections in an infants ward

Nosocomial viral respiratory infections: Perennial weeds on pediatric wards

Control of nosocomial respiratory syncytial viral infections

Nosocomial respiratory syncytial virus infection

Possible transmission by fomites of respiratory syncytial virus

Respiratory syncytial virus in adults

Respiratory syncytial virus infection in young hospitalized children. Identification of risk patients and prevention of nosocomial spread by rapid diagnosis

Restriction endonuclease fingerprinting of herpes simplex virus DNA: A novel epidemiological tool applied to a nosocomial outbreak

Herpes simplex virus infection of the adult lower respiratory tract

An outbreak of var-156

icella-zoster virus infection among cancer patients

Spread of varicella in hospitalized children having no direct contact with an indicator zoster case and its prevention by live vaccine

Live-attenuated varicella virus vaccine

Nosocomial conjunctivitis caused by adenovirus type 4

Epidemiology of cytomegalovirus infection in end stage renal disease

A nosocomial outbreak of influenza A

Parainfluenza outbreaks in extended care facilities: United States

Nosocomial influenza B virus infection in the elderly

Rhinovirus transmission: One if by air, two if by hand

Pneumocystis carinii pneumonia

Pneumocystis carinii pneumonia: A cluster of eleven cases

An outbreak of Pneumocystis pneumonia in children with acute leukemia

An outbreak of Pneumocystis carinii pneumonia at a pediatric hospital

Scabies: An epidemiologic reassessment

Hospital epidemic of scabies diagnosis and control

Nosocomial outbreak of scabies

To improve service to our subscribers, we will be changing the volume year of Disease a Month to coincide with the calendar year. In other words, January 1985 will be Volume XXXI, number 1. The October through December issues will be Volume XXX numbers 13, 14, and 15 respectively. This change will enable us to publish a cumulative annual index in the December issue which will also reference a complete volume year. Your subscription renewal date will not be affected in any way by this change.