Education + Advocacy = Change

Click a topic below for an index of articles:




Financial or Socio-Economic Issues


Health Insurance



Institutional Issues

International Reports

Legal Concerns

Math Models or Methods to Predict Trends

Medical Issues

Our Sponsors

Occupational Concerns

Our Board

Religion and infectious diseases

State Governments

Stigma or Discrimination Issues


If you would like to submit an article to this website, email us at for a review of this paper

any words all words
Results per page:

“The only thing necessary for these diseases to the triumph is for good people and governments to do nothing.”


Infections Acquired during Cardiopulmonary Resuscitation

Estimating the Risk and Defining Strategies for Prevention

George C. Mejicano, MD and Dennis G. Maki, MD

15 November 1998 | Volume 129 Issue 10 | Pages 813-828

Purpose: To estimate the risk for acquiring an infectious disease during cardiopulmonary resuscitation (CPR) or CPR training and to identify strategies to minimize that risk.

Data Sources: English-language articles published since 1965 were identified through a search of the MEDLINE database and selected bibliographies.

Study Selection: Studies that contained information about transmission of infectious organisms, particularly HIV and other bloodborne viruses that might be transmitted through mouth-to-mouth ventilation, contact exposures, and needlesticks during CPR.

Data Extraction: Descriptive and analytic data from each study.

Data Synthesis: Fear of acquiring infection, especially HIV infection, can delay prompt initiation of mouth-to-mouth ventilation. Although pathogens can be isolated from the saliva of infected persons, salivary transmission of bloodborne viruses is unusual and transmission of infection has been rare: Only 15 documented cases have been reported. Most of these cases involved a bacterial pathogen, such as Neisseria meningitidis. Transmission of hepatitis B virus, hepatitis C virus, or cytomegalovirus during CPR has not been reported; all three reported cases of HIV infection acquired during resuscitation of an infected patient resulted from high-risk cutaneous exposures. There have been no reports of infection acquired during CPR training. Simple infection-control measures, including use of barrier devices, can reduce the risk for acquisition of an infectious disease during CPR and CPR training. Postexposure protocols can further protect potential rescuers and trainees.

Conclusions: The benefit of initiating lifesaving resuscitation in a patient in cardiopulmonary arrest greatly outweighs the risk for secondary infection in the rescuer or the patient. Nevertheless, use of simple infection-control measures during CPR and CPR training can reduce a very low level of risk even further.

"And he went up, and lay upon the child, and put his mouth upon his mouth, and his eyes upon his eyes, and his hands upon his hands: and he stretched himself upon the child; and the flesh of the child waxed warm."

II Kings 4:34

From the time of this first written description of rescue breathing [1], it was nearly 3000 years before mouth-to-mouth ventilation achieved widespread medical acceptance [2]. The National Research Council first recommended mouth-to-mouth ventilation in 1957, and the first reported out-of-hospital resuscitation took place in 1960 [3]. Today, mouth-to-mouth ventilation combined with external chest compressions, termed cardiopulmonary resuscitation (CPR), form the initial treatment of patients with cardiac arrest. In the United States, an estimated 100 000 resuscitations are done annually, and more than 20 million citizens to date have been formally trained in basic life support [3].

Large-scale educational programs have put CPR at the forefront of public health initiatives on the basis of the knowledge that prompt action can save lives, both inside and outside of the hospital [4,5]. Average survival rates for patients with cardiac arrest are low, even with prompt administration of CPR-10% to 15% [6,7]-and factors such as age [4], location of arrest [4,5], and duration of arrest [4] profoundly influence the outcome of resuscitation. However, the strongest predictors of success are whether the arrest is witnessed [4,6,8], the initial cardiac rhythm [6,8,9], the time from collapse to initiation of CPR [8,10-12], and the quality of the CPR administered [9]. The highest rates of survival and hospital discharge, up to 43%, occur when CPR is started within 3 to 4 minutes of arrest [7,8,12]. When CPR is combined with immediate on-site defibrillation, the survival rate may be as great as 70% [11]. The most recent guidelines from the American Heart Association [13] emphasize that prompt intervention is the key to success: A first responder must be willing to act without hesitation. Unfortunately, fear of contracting a communicable disease, especially HIV infection, has become a major barrier to immediate response [14,15].

Impact of AIDS on Initiation of Cardiopulmonary Resuscitation

Most students in the health professions [16], nurses [17-19], and house officers [20,21] have stated in surveys that they would not perform mouth-to-mouth ventilation on a patient with AIDS. A recent survey [22] found that 45% of 433 internists and 80% of 152 nurses would not perform mouth-to-mouth ventilation on a stranger. The main reason for reluctance was fear of contracting a serious communicable disease, especially HIV infection.

Even CPR instructors harbor these fears. A survey of 1794 instructors in Virginia found that 49% had performed CPR within 3 years. Of these providers, 40% reported that they had hesitated to provide mouth-to-mouth ventilation at least once, usually because of fear of exposure to a contagious disease [23]. Not surprisingly, many instructors also have concerns about exposure to infectious diseases during group training done by using mannequins [23-25].

In 1989, the Emergency Cardiac Care Committee of the American Heart Association [26] sought to allay concerns about potential exposure to infectious agents, particularly HIV and hepatitis B virus (HBV). They stressed that "delayed ventilation could mean death or disablement for an otherwise healthy person, while risk to the rescuer, even with a known HBV/HIV-positive victim, is considered very low." An editorial and a position paper from the Heart and Stroke Foundation of Canada [27,28] reaffirmed this position. The current American Heart Association guidelines [26] state that health care workers and public protection professionals have "moral, ethical, and, in certain situations, legal obligations to provide CPR." We concur with these statements but believe that it will take more than reassurances and ethical arguments to persuade health care workers and the lay public that fear of infection should not be a deterrent to the initiation of CPR.

Although the absolute level of risk is low, health care workers have an increased risk for occupationally acquired infection, particularly infection with Mycobacterium tuberculosis and HBV [29-31]. Yet recent reviews of occupationally acquired infection in health care workers have not addressed the risk for infection as a consequence of administering CPR. By examining the available data on the likelihood of transmission of infectious organisms (particularly HIV and other ubiquitous bloodborne pathogens) during CPR and the documented cases of infections acquired during CPR and by summarizing essential infection-control measures that reduce risk even further, we believe that it is possible to convince potential CPR providers that the risk for acquiring an infection during CPR is not zero but is very low and that there should be no reluctance to begin CPR, including mouth-to-mouth ventilation, when the procedure is medically and ethically indicated.

Risk for Salivary Transmission of Bloodborne Viruses


It is now abundantly clear that HIV is not transmitted through casual contact [32-35] (Table 1). Even household members who have shared numerous personal items with HIV-infected persons, such as eating utensils, razors, or toothbrushes, have not become infected [34]. Infection with HIV is acquired through sexual contact; exposure to unscreened blood products; intravenous drug use; and, among health care workers, exposure to contaminated sharps. Infants acquire HIV in utero or during parturition [32,46]. Transmission by all other routes seems to be very rare [32,39].


Table 1. Risk for Salivary Transmission of HIV


Human immunodeficiency virus is isolated rarely, and then in very low concentrations, from the saliva of HIV-infected patients [47,48]. The low concentration of HIV in saliva is probably the major reason why dental professionals have a very low risk for occupationally acquired HIV infection [36,37]. In the United States, no rigorously documented cases of transmission of HIV from a patient to a dental worker have been reported, although seven cases may have derived from dental exposures [37]. In one study of 1309 dental professionals who had no behavioral risk factors for HIV infection but had cared for multiple patients known to have AIDS or risk factors for HIV infection [36], only 1 dentist was seropositive for HIV.

In a long-term prospective study of health care workers heavily exposed to patients with HIV infection [38], none of the 63 persons exposed to the saliva of HIV-positive patients on a daily basis developed antibody to the virus. Even persons who have been bitten by an HIV-infected person seem to be at low risk: Only four cases of HIV infection have been ascribed to human bites, and none of these cases have been confirmed epidemiologically by RNA subtyping [41-44]. Of 8 HIV-seronegative children bitten by HIV-infected playmates and 38 health care workers bitten by patients with AIDS, none have seroconverted despite prolonged followup [39,40,49]. Moreover, two nurses who performed mouth-to-mouth ventilation on a patient with AIDS have not developed HIV infection during prolonged follow-up [45].

In summary, although there has been speculation that salivary transmission of HIV may occur, the evidence does not support this [32,41,50,51] (Table 1).

Hepatitis B Virus

In contrast to HIV, HBV poses substantial risks to health care workers [41,52-54]. Hepatitis B surface antigen (HBsAg) was found in the saliva of 31 of 41 patients (76%) with acute hepatitis and was found intermittently in 75 of 93 patients (81%) with chronic HBV infection [55]. Thus, salivary exchange may be one mechanism for nonparenteral transmission of HBV within families [56-58], and cases of HBV infection acquired through human bites have been reported [59,60] (Table 2).



Table 2. Risk for Salivary Transmission of Hepatitis B Virus



However, the sum total of the evidence suggests that salivary transmission of HBV is rare (Table 2). Only one case of HBV infection was found after 35 upper gastrointestinal procedures were done with endoscopes that were presumably contaminated with HBV [61,62]. Moreover, none of 12 students who were exposed to HBsAg-positive saliva through the sharing of musical instruments with an infected teacher seroconverted or developed clinical evidence of hepatitis [63]. Finally, none of 39 persons exposed to HBsAg-positive saliva during CPR training became infected [63,64], and no cases of HBV infection acquired as a consequence of giving CPR or participating in CPR training have been documented. The most likely reason why salivary transmission is so infrequent is that the concentration of HBV in saliva is only 1/3000 of that in serum [65]. The risk for acquiring HBV infection as a result of performing mouth-to-mouth ventilation is very low and can almost certainly be mitigated by immunization of the health care workers and public protection officials most likely to be called on regularly to perform CPR [66].

Hepatitis C Virus

Unlike patients infected with HIV and HBV, up to one third of persons with antibody to the hepatitis C virus (HCV) have no identifiable risk factors for bloodborne infection [67,68]. Hepatitis C virus RNA has been found by polymerase chain reaction (PCR) in many body fluids, including sweat [69], urine [68-70], and saliva [68-76]. The prevalence of HCV positivity in saliva has ranged widely, from none of 14 HCV-seropositive patients in one study [77] to 20% to 62% of HCV-infected patients in others [70,74,75]; this suggests wide differences in the sensitivity of the assays used to assess positivity. It seems clear, however, that the concentration of HCV in blood is several orders of magnitude greater than the concentration of HCV in saliva [70].

In one study (Table 3), none of 62 household contacts of HCV-seropositive family members showed serologic evidence of infection [78]; in a prospective study of the spouses of 11 patients with active HCV infection [76], none were positive for HCV RNA by PCR. Similarly, other investigators have found low rates (1.4%) of HCV seropositivity in nonsexual household contacts of HCV-positive persons [79].



Table 3. Risk for Salivary Transmission of Hepatitis C Virus




Despite the frequent presence of HCV in saliva, only 2 of 305 general dentists (0.7%) and 7 of 343 oral surgeons (2.0%) had serologic evidence of HCV infection, although 14.7% of both groups showed one or more markers for HBV infection [54]. Only a single case of HCV infection has been attributed to salivary transmission, and this case was in the recipient of a human bite [82].

At this time, saliva does not seem to play a major role in transmission of HCV, although more work is needed to better elucidate how HCV is transmitted from person to person (Table 3). It can be concluded that the risk for acquiring HCV through the performance of CPR is very low.


Cytomegalovirus (CMV) is a teratogenic virus that has attracted considerable attention as a potential threat to female health care workers [83]. Outbreaks of CMV infection in day care centers and the survival of CMV on fomites, such as paper diapers [84], suggest that this ubiquitous virus is more easily transmitted through nonparenteral than through parenteral routes. Cytomegalovirus can be found in saliva, both by culture and by PCR, and co-infection with HIV increases the likelihood of finding CMV in saliva [85-87].

Many cohort studies in health care workers have assessed the occupational risk for acquiring CMV (Table 4). Seven studies [88-94] produced remarkably similar results: a mean annual seroconversion rate among heavily exposed health care workers of 3%. Despite frequent exposure to infants and children shedding CMV, the risk of health care workers for acquisition of CMV was not much higher than the risk of community controls (mean, 1.6%) [90,93,96] and was much lower than the risk of day care workers (mean, 14%) [95,96]. A recent meta-analysis [97] concluded that pediatric nurses are not at increased risk for occupationally acquired CMV infection. Finally, a seronegative pediatric house officer who administered mouth-to-mouth ventilation to a newborn with congenital CMV infection did not acquire CMV, despite ingestion of the infant's CMV-positive secretions [98].



Table 4. Risk for Non-Bloodborne Transmission of Cytomegalovirus

Table 4. Risk for Non-Bloodborne Transmission of Cytomegalovirus



Young women who work with CMV-infected persons must be apprised that they are at some risk (although this risk is low) for infection with CMV. Compliance with handwashing can reduce this risk. Pregnant health care workers and their sexual partners can safely care for CMV-infected patients if they follow universal precautions [97,99-103].

Data that would pinpoint the risk for salivary transmission of CMV are, unfortunately, not available. However, despite the ubiquity of CMV and its frequent presence in saliva, the information we have indicates that the risk for acquiring CMV through the administration of mouth-to-mouth ventilation is probably not great.

Creutzfeldt-Jakob Agent

Creutzfeldt-Jakob disease, which is rare in humans, is one of several devastating neurologic diseases caused by prions (proteinaceous and infectious agents) [104-106]. Recently, the emergence of a new variant of Creutzfeldt-Jakob disease in the United Kingdom, circumstantially linked to bovine spongiform encephalopathy ("mad cow disease"), has attracted attention [105,106]. Rare clusters of Creutzfeldt-Jakob disease have been seen in some ethnic groups [107], and cases of suspected iatrogenic transmission have been linked to corneal transplantation [108], intracerebral electroencephalography [109], dura mater implantation [110-112], and exposure to cadaveric growth hormone [113,114]. Isolated case reports have implicated blood transfusion [115] and solid organ transplantation [116]. Primates inoculated with tissue from the spinal cord, eye, lung, liver, kidney, and spleen of persons with Creutzfeldt-Jakob disease acquired the disease, but they did not acquire it through exposure to the saliva or blood of infected persons [114]. Indeed, no cases of Creutzfeld-Jakob disease resulting from exposure to saliva have been documented. As of 1994, 24 cases of putative acquisition of Creutzfeldt-Jakob disease by health care workers through occupational exposure had been described [117]. However, the reported incidence of Creutzfeldt-Jakob disease is no higher in health care workers than in the general population [118].

Given the rarity of Creutzfeldt-Jakob disease and the lack of clinical or experimental data showing the plausibility of salivary transmission, it is unlikely that a rescuer performing mouth-to-mouth ventilation could contract this disease as a result of efforts to resuscitate a person in cardiopulmonary arrest.

Risk for Transmission of Bloodborne Viruses through Parenteral Exposure

Although the likelihood of acquiring a bloodborne virus, such as HIV, HBV, or HCV, from exposure to saliva is extremely low (Table 1, Table 2, and Table 3), the risk for acquiring such a virus through a parenteral biohazardous exposure is substantial (Table 5) [119-127]. In general, the risk for transmission of these viruses depends on the extent and the circumstances of exposure. With HIV, deep, penetrating injuries, especially with hollow needles, greatly increase risk; moreover, the higher the concentration of HIV in the source patient's blood, the greater the likelihood of transmission (Table 6) [128]. With HBV, the presence of detectable e antigen (HBeAg), which is associated with a much higher concentration of HBV in the blood, greatly increases risk: The likelihood of transmission of HBV is at least 10-fold greater with parenteral exposure to HBeAg-positive blood than with exposure to blood from an HBsAg-positive, HBeAg-negative person (Table 5) [123-125].



Table 5. Risks for Transmission of Bloodborne Viruses by Parenteral Exposure





Table 6. Risk Factors for Transmission of HIV after Percutaneous Exposure to HIV-Infected Blood*






In general (Table 5), the risk resulting from parenteral exposure is greatest with HBV (13.1%) [123-125], intermediate with HCV (5.8%) [126,127], and lowest with HIV (0.32%) [119-122].

Infections That Have Been Acquired through Cardiopulmonary Resuscitation

During the past 30 years, CPR has been given to millions of persons worldwide [3]. Thus, it is remarkable that only 15 infections convincingly linked to resuscitation have been reported (almost all from the United States) (Table 7) [129-139]. With an estimated 100 000 resuscitations performed in the United States each year [3], this suggests a risk for infection of less than 1 in 200 000. However, it must be acknowledged that underreporting undoubtedly occurs and that the actual number of infections is almost certainly greater. Even if it were 100-fold greater, the absolute level of risk would still be less than 1 in 2000.


Table 7. Infectious Organisms Transmitted during Cardiopulmonary Resuscitation



Mouth-to-Mouth Transmission

Bloodborne Viruses

No cases of HIV, HBV, HCV, or CMV infection transmitted by mouth-to-mouth ventilation have been documented. Bierens and Berden [141] estimate that the risk for acquiring HIV infection during mouth-to-mouth ventilation is between 1 in one million and 1 in one billion, depending on the population being resuscitated and the circumstances of the exposure.

Mycobacterium tuberculosis

The earliest report of a communicable disease clearly acquired as a consequence of administering CPR was in 1965: A medical intern developed primary cutaneous tuberculosis 8 weeks after giving mouth-to-mouth ventilation to a patient in cardiac arrest who was later found, at autopsy, to have active pulmonary tuberculosis [129]. Despite the large number of persons with active tuberculosis worldwide, no other reported cases are thought to have been acquired during CPR. Still, a rescuer who has given mouth-to-mouth ventilation to a patient with pulmonary tuberculosis should have serial tuberculin skin testing and should be offered chemoprophylaxis if conversion occurs [142].


Neisseria meningitidis

At least four cases of meningococcal infection in health care workers have been linked to rescue breathing [130]. The risk for salivary transmission may be high during mouth-to-mouth ventilation or another intensive respiratory exposure (such as endotracheal intubation) if a patient has systemic N. meningitidis infection [143,144]. Providers who have such an exposure should be offered chemoprophylaxis [143-146].

Enteric Pathogens

There have been rare cases of transmission of gastrointestinal pathogens during rescue breathing. A physician developed fever and diarrhea 3 days after performing mouth-to-mouth ventilation on a moribund infant; cultures from the physician and the child showed Shigella sonnei [131]. Another report documented transmission of Salmonella infantis from a 68-year-old woman to the physician who had attempted to resuscitate her with mouth-to-mouth ventilation [132]. A recent case report [147] suggests that Helicobacter pylori was transmitted, through mouth-to-mouth ventilation, but definitive evidence of this is lacking.

Herpes Simplex Virus

Herpes simplex virus (HSV) has been shown to be readily transmitted both from patient to rescuer and from rescuer to patient [135-137], reflecting the ubiquity of HSV infection in the general population [148]. Viable HSV has been obtained from the saliva of 2.5% of asymptomatic adults [149], and copious shedding of HSV occurs in persons with active lesions: One study [150] reported the isolation of HSV from 78% of salivary samples and 67% of hand cultures from persons with active lesions.

Given these statistics, it is surprising that only two cases of primary herpes labialis secondary to CPR have been reported; both involved medical residents who performed mouth-to-mouth ventilation [135,136]. Herpes simplex virus was also transmitted from a physician to a patient during CPR [137] (see below).

Other Respiratory Pathogens

In a survey of 27 pediatric house officers, 74% reported ingesting an infant's secretions during neonatal resuscitation [133]. Eight residents reported exposure to culture-proven respiratory pathogens during neonatal suctioning; in one case, cultures from the house officer matched those from the infant (Neisseria gonorrhoeae). The authors of this study [133] estimated that the risk for ingesting secretions that contain a pathogen during neonatal resuscitation is 1 in 300.

Contact Transmission

Group A Streptococci

A firefighter ventilated a moribund 3-year-old boy with a bag-valve-mask device but contaminated his hands with the child's secretions during the resuscitation. While cleaning his equipment afterward, the firefighter sustained a small abrasion. Within 24 hours, he became acutely ill with cellulitis emanating from the abrasion, shock, and acute renal failure. Cultures from the firefighter's hand wound and from the resuscitated child's blood grew a toxigenic strain of Streptococcus pyogenes [134].


In 1986, a nurse assisted in the resuscitation of a young man with undiagnosed AIDS, heavily contaminating her ungloved hands with the patient's blood during insertion of an arterial catheter. Four months later, the nurse was found to be HIV positive. This is one of three well-documented cases in which HIV infection seems to have been transmitted through contamination of skin by blood [139]; in each case, the exposure was heavy and prolonged and the exposed health care worker had chapped hands or other possible skin breaks. These cases form the basis for the central role of gloves in universal precautions. It is essential to reemphasize, however, that only one case of occupationally acquired HIV infection was seen among 2071 prospectively studied health care workers who had a mucous membrane or non-intact skin exposure to HIV-positive blood [151].

Needlestick Transmission

The scene of a resuscitation is often chaotic, and health care workers engaged in CPR are therefore at risk for exposure to bloodborne pathogens from accidental needlesticks and other sharps injuries. Pooled data from 25 prospective studies of health care workers exposed to HIV through needlesticks or other parenteral exposures have quantified the risk for acquiring HIV as 0.32% (95% CI, 0.18% to 0.46%) (21 infections after 6498 exposures) [119-122,151-156]. Two of the cases occurred during CPR and involved deep needlesticks [138,140]. It is likely that additional cases of occupationally acquired HIV infection (among the 163 to 177 cases reported in the United States as of 1996 [151,157]) derived from sharps exposures during CPR, but the circumstances of the exposures were not reported. Similarly, it is almost certain that HBV and HCV infections have originated from parenteral exposures to these viruses during administration of CPR to infected patients, but no such cases have been reported to date.

Prevention of sharps injuries is essential because most exposures to bloodborne viruses can easily be avoided-fully 37% in the Centers for Disease Control and Prevention (CDC) HIV Surveillance Study [138]. This is also true outside of the hospital setting. A study of occupational exposures in emergency medical service workers [158] found that most needlesticks occurred during needle disposal.

Eliminating recapping of needles, making impervious containers for sharps disposal widely available, achieving maximal compliance with proper sharps disposal guidelines, and increasing use of needleless systems and other safety devices form the cornerstone of an institutional program aimed at reducing the risks of bloodborne pathogens to health care workers (Table 8) [159-161].


Transmission from Rescuer to Patient

Concerns about the safety of CPR usually focus on risk to the rescuer, but the potential risk posed to a patient by an infected rescuer also warrants consideration. However, we know of only one case in which a patient acquired an infection as a result of contact with an infected rescuer: A pediatric resident was shown by DNA subtyping to have transmitted HSV from a crusting cold sore on his lip to a neonate whom he was resuscitating through an endotracheal tube [137].

In clinical practice, if rescuers have open sores on the hands or the face or lesions in the mouth, use of a barrier device is strongly recommended [13]. However, if no barrier devices are immediately available, the imperative to quickly initiate life-sustaining measures, including mouth-to-mouth ventilation, must take precedence over concern about transmission of an infectious agent [13].

The best evidence suggests that although the potential exists for acquisition of infection by a patient during mouth-to-mouth ventilation, the risk is almost certainly very low (Table 1, Table 2, Table 3, and Table 4). Because expeditious resuscitation is crucial, it is not only acceptable but desirable that a qualified rescuer begin resuscitory efforts, including mouth-to-mouth ventilation, even if the rescuer is known to be positive for HIV, HBsAg, or HCV [13,26]. Obviously, if an oral barrier (such as a pocket mask or bag-valve-mask) is available, it should be used.

If unprotected mouth-to-mouth ventilation has been given and the recipient survives, the recipient must be thoroughly evaluated for infection acquired through CPR (Table 8).

Precautions To Reduce Risk during Cardiopulmonary Resuscitation

An Institutional Program

The Occupational Safety and Health Administration mandates that health care institutions establish programs to protect health care workers from biohazardous exposures, especially exposures to infectious diseases [162-164]. These programs must include education of all employees with respect to risk reduction; a multifaceted strategy for prevention of sharps injuries and other biohazardous exposures; and a postexposure protocol to expeditiously determine the potential risk associated with an individual exposure and, when indicated, to provide prophylaxis and long-term follow-up. Training in CPR must address protection from infection during CPR, with the greatest focus on measures for avoiding sharps injuries. Table 8 provides model institutional and general guidelines for prevention and management of the transmission of infectious agents during CPR.

Universal Precautions

Health care workers cannot clinically identify most patients infected with bloodborne pathogens [165,166]; thus, in 1987, the CDC published recommendations urging the use of "universal precautions" with all patients. These recommendations were based on the premise that every patient must, as a precaution, be considered HIV positive [167]. An update published in 1988 [168] stressed that blood and serum pose the greatest risks for transmission of HBV and HIV and that "general infection control practices already in existence-including the use of gloves for digital examination of mucous membranes and endotracheal suctioning, and handwashing after exposure to saliva-should further minimize the minute risk, if any, for salivary transmission of HIV and HBV."

Oral Barrier Devices

Despite the CDC's exclusion of saliva as a high-risk body fluid with respect to universal precautions [168], support for measures to avoid or minimize contact with saliva during mouth-to-mouth ventilation and CPR has been almost unanimous. The University of California-San Francisco, Task Force on the Acquired Immunodeficiency Syndrome [169], the American Hospital Association [170], the American College of Emergency Physicians [171], and the U.S. Public Health Service [172] have endorsed the use of disposable devices (Figure 1), pocket masks, and bag-valve-masks to prevent direct mouth-to-mouth contact between rescuer and patient.


Figure 1. Oral barrier devices. Top. Adult face mask (Adult Classic Face Mask with Adjustable Cushion, Vital Signs, Inc., Totowa, New Jersey) with one-way valve (Vent Easy II Non-rebreathing Valve, Respironics, Inc., Murrysville, Pennsylvania). Middle. Adult face mask with one-way valve and extension (Carhill Valve Resuscitation Device, Bird Life Design, Dallas, Texas). Bottom. Laerdal bag-valve-mask (Vital Signs, Inc.).


Not all barrier devices are equally efficacious [141]. When 17 mask and face-shield resuscitation devices were evaluated in a 60-second simulation of rescue breathing [173], most mask devices prevented transmission of oral organisms but 6 of 8 face shields were contaminated [173]. All ventilation devices with a one-way valve (with the exception of one brand) seemed to prevent transmission from the simulated patient's oral flora to the rescuer's side of the device. Another device was tested in a simulation model with an aerosol containing Staphylococcus aureus: At best, only 90% of organisms were retained [174]. It is unclear how far these data can be extrapolated to real-life conditions [175].

Which type of device optimizes external ventilation? When working with a bag-valve-mask device, using two hands improves the volume of air delivered to a patient [176]. Because two hands are needed to squeeze the bag properly, another rescuer is needed to ensure a tight fit between the mask and the patient-that is, three rescuers are better than two [176-178]. Because of this, some experts have advocated the use of mouth-to-mask devices [179,180]. However, mouth-to-mask methods seem to be second in effectiveness to mouth-to-mouth methods [181]. One study [181] evaluated 35 providers to determine how effectively they "ventilated" a mannequin. After instruction, effective ventilation was achieved by 97% of the providers who used mouth-to-mouth and mouth-to-mask methods. Bag-valve-mask ventilation, however, was suboptimal 97% of the time [181].

Another problem with bag-valve-masks and mouth-to-mask devices is that they are not always readily available, even in hospitals. One study that reviewed the results of 38 cardiopulmonary arrests in a tertiary care hospital found substantial delays in the initiation of rescue breathing. Resuscitation was performed within 1 minute in only 37% of resuscitations; initiation took longer than 3 minutes in 18% of resuscitations. Many of the delays occurred because barrier devices were not immediately available and the rescuers chose not to give mouth-to-mouth ventilation [182].

Despite all of these concerns, the adoption of barrier devices has become de facto policy in the health care setting, especially in the public protection sector. New York City requires specified public places, including bars, theaters, health clubs, and restaurants, to have infection-control resuscitation equipment available [183]. A study found that having protective materials readily available was probably more important than education alone in improving compliance with the use of barrier precautions [184].

Management of Exposures to Infectious Diseases during Cardiopulmonary Resuscitation

Indications for Evaluation after Cardiopulmonary Resuscitation

It is essential that persons performing CPR have access to a postexposure protocol (Table 8). This protocol will be most consistently and conveniently administered through an institutional employee health department [159].

After CPR, especially if mouth-to-mouth resuscitation has been given, a formal effort should be made to ascertain whether the patient could have had a serious, contagious respiratory infection. This can usually be done by reviewing the available clinical history and the results of routine diagnostic studies, such as screening blood tests and chest radiography. If a biohazardous exposure has occurred (if blood or open sores were apparent in the patient's mouth and mouth-to-mouth ventilation was performed or if a member of the resuscitation team sustained a needlestick or a mucous membrane or non-intact skin exposure to the patient's blood) the patient should be tested for infection with bloodborne viruses (Table 8) [159]. If CPR is unsuccessful, needed cultures and blood specimens must be obtained before the patient's body is released for embalming or cremation.

Evaluation of the Source Patient

In the event of a biohazardous exposure during CPR, the source patient's history must be ascertained as quickly as possible to determine the presence of communicable disease and the magnitude of risk (Table 8) [159]. The source patient's blood should be tested for evidence of infection with HIV, HBV, and HCV [122,151,185]; HIV testing must be done in compliance with state statutes. With respiratory contact, especially unprotected mouth-to-mouth ventilation, an effort should be made to determine whether the source patient is likely to have active N. meningitidis or M. tuberculosis infection. If possible, the source patient should be examined for evidence of active infection, such as labial HSV infection. If clinical and epidemiologic data from the source patient are unavailable, the source patient's risk for harboring a bloodborne virus should be estimated to guide follow-up care of the exposed health care worker [122].

Management of the Exposed Rescuer

An exposed rescuer should immediately and copiously rinse any puncture wound or exposed mucous membrane with tap water (Table 8) [159]. If the exposure involves a sharps injury, the area should be squeezed to induce bleeding and should then undergo surface disinfection with a virucidal agent, such as an iodophor antiseptic [159]. Baseline studies in the exposed rescuer should include serologic tests for infection with HIV, HBV, and HCV [122,185]. The exposed rescuer should also be counseled about the potential risks for acquiring an infectious disease.

Agreement on the management of exposures to N. meningitidis [143-146] and M. tuberculosis is clear-cut [142] (Table 9). A rescuer exposed to N. meningitidis should receive prophylaxis with rifampin, ciprofloxacin, or (if the rescuer is pregnant) ceftriaxone [146]. If the source patient is found to have active pulmonary tuberculosis, baseline and follow-up tuberculin tests at 12 weeks are indicated. After mouth-to-mouth or other close respiratory tract exposure, isoniazid chemoprophylaxis should be started immediately unless the rescuer is immunocompetent and known to be tuberculin-positive. Chemoprophylaxis should be continued for 6 months if tuberculin conversion is documented; if the rescuer is immunocompromised, chemoprophylaxis should be continued regardless of the rescuer's baseline tuberculin status. If the exposure involves isoniazid-resistant disease, rifampin should be used in place of isoniazid. For exposure to multidrug-resistant tuberculosis, pyrazinamide plus a quinolone or ethambutol is recommended [142].



Prophylactic measures to protect rescuers exposed to blood or body fluids that harbor HBV are also well defined [66,122]. If the rescuer has not been immunized against HBV, one dose of HBV hyperimmune globulin (0.06 mL/kg of body weight intramuscularly) should be given immediately and the three-dose HBV vaccine series should be initiated. If the exposed rescuer has been immunized, measurement of the rescuer's level of antibody to HBV will dictate whether HBV hyperimmune globulin, a booster dose of the vaccine, or both should be given [66,122].

In contrast, because donated blood containing detectable antibody to HCV is now excluded in the preparation of standard immune globulin [122], the Advisory Committee on Immunization Practices has concluded that there is "no basis for supporting the use of immune globulin for post-exposure prophylaxis of hepatitis C." Furthermore, the use of other antiviral agents (such as interferon-alpha) in postexposure prophylaxis against hepatitis C is not recommended [187].

Until recently [188], the value of postexposure prophylaxis against HIV with antiretroviral drugs was controversial; in part, it was related to at least 14 cases of failure of zidovudine prophylaxis to prevent infection in health care workers after percutaneous HIV exposure [153,186]. However, a recent international case-control study (Table 6) found that postexposure use of zidovudine was associated with a lower risk for HIV transmission (adjusted odds ratio, 0.2 [CI, 0.1 to 0.6]) [128,189]. Moreover, the combination of purine analogues and protease inhibitors may be superior to zidovudine alone [119,190] and seems to be cost-effective [191]. For these reasons, in May 1998, the U.S. Public Health Service issued updated guidelines for antiretroviral prophylaxis after occupational exposure to HIV; in these guidelines, the antiretroviral regimen recommended is based on a detailed algorithm for scoring the severity of the exposure and the estimated magnitude of risk [186]. For high-risk percutaneous exposures to HIV (such as a deep injury caused by a hollow needle), a 4-week, three-drug regimen with zidovudine, lamivudine, and indinavir or nelfinavir should be considered (Table 9).

With mucosal contact, the benefit of postexposure prophylaxis with antiretroviral drugs remains controversial [122,188], and the new guidelines do not advocate use of these drugs unless saliva has been contaminated by HIV-positive blood [186]. Prophylaxis is not recommended for cases of cutaneous contact unless the exposure involves a high viral titer, contact is prolonged, the involved area is extensive, or skin integrity is visibly compromised [186].

Exposed rescuers must receive close follow-up [159] and, if exposed to bloodborne viruses, must be retested for evidence of infection for at least 6 to 9 months after the exposure [122,186]. The CDC recommends that health care workers who are seronegative for HIV immediately after exposure to HIV should be retested 6 weeks, 12 weeks, and 6 months after exposure [186,188,192].

Measures To Reduce Risk during Training for Cardiopulmonary Resuscitation

Mannequin Disinfection

Although the spread of infection through the use of training mannequins has more in common with transmission through fomites or endoscopes than with transmission through contact with living patients, the use of precautions has been extended to CPR training since the possibility that mannequins might harbor infectious agents has been raised [193-195]. The CDC and the American Heart Association have issued guidelines on how to clean and disinfect mannequins [196-198]; following these guidelines can make CPR training safer and allay the fears of would-be CPR providers [199-202].

Training the Infected Rescuer

It seems prudent to postpone CPR training if a person is acutely ill with a potentially contagious disease that can be transmitted through respiratory secretions (such as a cold, influenza, or herpes labialis) or has an open lesion or lesions on the face or hands [26,198,203-205].

A potential CPR trainee with a chronic infection, such as HIV or HBV infection, is another matter. The American Heart Association addressed this issue in their most recent Instructor's Manual [198]. Because speed is of the essence in a cardiopulmonary emergency, exclusion of a potential rescuer with a known chronic infection might preclude a patient's chance of survival. Thus, with rare exceptions, anyone who desires CPR training can receive it, including those who are positive for HIV, HBsAg, or HCV, provided that published precautions-use of a separate mannequin and recommended disinfection at the end of the class-are followed stringently [198].




Every health care provider tacitly accepts that the privilege of healing and comforting the sick is associated with some risk for exposure to contagious disease. However, the provider is not expected to take undue risks, and every effort must be made to reduce risk.

On the basis of the data reviewed here, we believe that the risk for acquiring an infectious disease during CPR or CPR training is low: Only 15 well-documented cases have been reported in the past 30 years. In contrast, the performance of CPR may be associated with a far higher risk for stress-related illness, such as myocardial infarction [206,207]. But even with allowance for 100-fold underreporting, the risk to a potential rescuer is still low, less than 1 in 2000 resuscitations; moreover, the risk for acquiring HIV, HBV, or HCV infection is extremely low, about 1 in one million [141]. In contrast, prompt performance of CPR can save the life of the person in cardiac arrest 10% to 15% of the time [6,7]. It seems clear that the societal benefit of administering CPR to a patient in cardiopulmonary arrest (10 000 to 15 000 lives saved per year in the United States) vastly outweighs the risks for secondary infection in the rescuer or the patient [205].

Recent research suggests that chest compression is more important in achieving adequate ventilation than mouth-to-mouth ventilation alone [208-211]; in the future, rescue breathing may not be considered necessary for basic cardiac life support. For now, it is still considered an integral feature of CPR [212].

Despite the very low risk for acquiring an infectious disease during CPR, efforts must be made to reduce risk even further (Table 8 and Table 9). Although more data are needed on the efficacy of oral barrier devices in preventing infection, these devices should be made more widely available, both in the hospital and in the community. Similarly, needleless systems should be used whenever possible, and procedures for the safe disposal of sharps must be taught and reemphasized. In addition, all susceptible rescuers and public protection professionals should be immunized against HBV. Training programs for CPR must emphasize personal protection and must follow recommended protocols for decontaminating mannequins that could harbor infectious agents.

Finally, every hospital must have a comprehensive protocol (Table 8) to assure expeditious evaluation and, when indicated, postexposure prophylaxis, counseling, and long-term follow-up. Public protection personnel, health care workers, and laypersons who perform CPR must also have access to such protocols, postexposure evaluation, and care.

Note: A longer version of this monograph will be published in Weil MH, Tang W, eds. CPR: Resuscitation of the Arrested Heart. Philadelphia: WB Saunders; [In press].

Grant Support: In part by an unrestricted grant for research in the prevention of infection from the Oscar Rennebohm Foundation.

Requests for Reprints: Dennis G. Maki, MD, H4/574 University of Wisconsin Hospital and Clinics, 600 Highland Avenue, Madison, WI 53792.

Current Author Addresses: Dr. Mejicano: J5/210 University of Wisconsin Hospital and Clinics, 600 Highland Avenue, Madison, WI 53792.

Dr. Maki: H5/574 University of Wisconsin Hospital and Clinics, 600 Highland Avenue, Madison, WI 53792.

Author and Article Information

From University of Wisconsin Medical School, University of Wisconsin at Madison, and University of Wisconsin Hospital and Clinics, Madison, Wisconsin.



1. II Kings 4:34. (Authorized King James version).

2. Comroe JH Jr. "... In comes the good air." Part II. Mouth-to-mouth method. Am Rev Respir Dis. 1979;119:1025-31.

3. Michael AD, Forrester JS. Mouth-to-mouth ventilation: the dying art. Am J Emerg Med. 1992;10:156-61

4. Tunstall-Pedoe H, Bailey L, Chamberlain DA, Marsden AK, Ward ME, Zideman DA. Survey of 3765 cardiopulmonary resuscitations in British hospitals (the BRESUS study): methods and overall results. BMJ. 1992;304:1347-51.

5. Litwin PE, Eisenberg MS, Hallstrom AP, Cummins RO. The location of collapse and its effect on survival from cardiac arrest. Ann Emerg Med. 1987;16:787-91.

6. Ballew KA. Cardiopulmonary resuscitation. BMJ. 1997;314:1462-5.

7. Lee KH, Angus DC, Abramson NS. Cardiopulmonary resuscitation: what cost to cheat death? Crit Care Med. 1996;24:2046-52.

8. Weston CF, Wilson RJ, Jones SD. Predicting survival from out-of-hospital cardiac arrest: a multivariate analysis. Resuscitation. 1997;34:27-34.

9. Gallagher EJ, Lombardi G, Gennis P. Effectiveness of bystander cardiopulmonary resuscitation and survival following out-of-hospital cardiac arrest. JAMA. 1995;274:1922-5.

10. Guidelines for cardiopulmonary resuscitation and emergency cardiac care. Emergency Cardiac Care Committee and Subcommittees, American Heart Association. Part I. Introduction. JAMA. 1992;268:2171-83.

11. Weaver WD, Cobb LA, Hallstrom AP, Copass MK, Ray R, Emery M, et al. Considerations for improving survival from out-of-hospital cardiac arrest. Ann Emerg Med. 1986;15:1181-6.

12. Eisenberg MS, Bergner L, Hallstrom A. Cardiac resuscitation in the community. Importance of rapid provision and implications for program planning. JAMA. 1979;241:1905-7.

13. Guidelines for cardiopulmonary resuscitation and emergency cardiac care. Emergency Cardiac Care Committee and Subcommittees, American Heart Association. Part II. Adult basic life support. JAMA. 1992;268:2184-98.

14. McNeil EL. Re-evaluation of cardiopulmonary resuscitation. Resuscitation. 1989;18:1-5.

15. Locke CJ, Berg RA, Sanders AB, Davis MF, Milander MM, Kern KB, et al. Bystander cardiopulmonary resuscitation. Concerns about mouth-to-mouth contact. Arch Intern Med. 1995;155:938-43.

16. Currey CJ, Johnson M, Ogden B. Willingness of health-professions students to treat patients with AIDS. Acad Med. 1990;65:472-4.

17. Sheard T. Going face to face with fear. Nursing. 1990;20:43.

18. Blumenfield M, Smith PJ, Milazzo J, Seropian S, Wormser GP. Survey of attitudes of nurses working with AIDS patients. Gen Hosp Psychiatry. 1987;9:58-63.

19. What our readers said about resuscitating a patient with A.I.D.S. Nurs Life. 1986;6:23-5.

20. Brenner B, Kauffman J, Sachter JJ. Comparison of the reluctance of house staff of metropolitan and suburban hospitals to perform mouth-to-mouth resuscitation. Resuscitation. 1996;32:5-12.

21. Brenner B, Stark B, Kauffman J. The reluctance of house staff to perform mouth-to-mouth resuscitation in the inpatient setting: what are the considerations? Resuscitation. 1994;28:185-93.

22. Brenner BE, Kauffman J. Reluctance of internists and medical nurses to perform mouth-to-mouth resuscitation. Arch Intern Med. 1993;153:1763-9.

23. Ornato JP, Hallagan LF, McMahan SB, Peeples EH, Rostafinski AG. Attitudes of BCLS instructors about mouth-to-mouth resuscitation during the AIDS epidemic. Ann Emerg Med. 1990;19:151-6.

24. Prosser RL Jr. Cardiopulmonary resuscitation training and risk of infection [Letter]. JAMA. 1990;263:3025.

25. Ornato JP. Providing CPR and emergency care during the AIDS epidemic. Journal of Emergency Care and Transportation: Emergency Medical Services. 1989;18:45-50.

26. Risk of infection during CPR training and rescue: supplemental guidelines. The Emergency Cardiac Care Committee of the American Heart Association. JAMA. 1989;262:2714-5.

27. Cummins RD. Infection control guidelines for CPR providers [Editorial]. JAMA. 1989;262:2732-3.

28. Mason M, Swanson RW, Warnica JW. Position statement. Disease transmission and cardio-pulmonary resuscitation. Emergency Cardiac Care Committee. Heart and Stroke Foundation of Canada. Canadian Journal of Cardiovascular Nursing. 1991;2:23-6.

29. Sepkowitz KA. Occupationally acquired infections in health care workers. Part I. Ann Intern Med. 1996;125:826-34.

30. Sepkowitz KA. Occupationally acquired infections in health care workers. Part II. Ann Intern Med. 1996;125:917-28.

31. Weiss SH. Risks and issues for the health care worker in the human immunodeficiency virus era. Med Clin North Am. 1997;81:555-75.

32. Friedland GH, Klein RS. Transmission, of the human immunodeficiency virus. N Engl J Med. 1987;317:1125-35.

33. Sande MA. Transmission of AIDS. The case against casual contagion [Editorial]. N Engl J Med. 1986;314:380-2.

34. Friedland GH, Saltzman BR, Rogers MF, Kahl PA, Lesser ML, Mayers MM, et al. Lack of transmission of HTLV-III/LAV infection to household contacts of patients with AIDS or AIDS-related complex with oral candidiasis. N Engl J Med. 1986;314:344-9.

35. Acquired immune deficiency syndrome (AIDS): precautions for clinical and laboratory staffs. MMWR Morb Mortal Wkly Rep. 1982;31:577-80.

36. Klein RS, Phelan JA, Freeman K, Schable C, Friedland GH, Trieger N, et al. Low occupational risk of human immunodeficiency virus infection among dental professionals. N Engl J Med. 1988;318:86-90.

37. Robert LM, Bell DM. HIV transmission in the health-care setting. Risks to health-care workers and patients. Infect Dis Clin North Am. 1994;8:319-29.

38. Gerberding JL, Bryant-LeBlanc CE, Nelson K, Moss AR, Osmond D, Chambers HF, et al. Risk of transmitting the human immunodeficiency virus, cytomegalovirus, and hepatitis B virus to health care workers exposed to patients with AIDS and AIDS-related conditions. J Infect Dis. 1987;156:1-8.

39. Lifson AR. Do alternate modes for transmission of human immunodeficiency virus exist? JAMA. 1988;259:1353-6.

40. Tsoukas C, Hadjis T, Shuster J, Theberge L, Feorino P, O'Shaughnessy M. Lack of transmission of HIV through human bites and scratches. J Acquir Immune Defic Syndr. 1988;1:505-7.

41. Richman KM, Rickman LS. The potential for transmission of human immunodeficiency virus through human bites. J Acquir Immune Defic Syndr. 1993;6:402-6.

42. Wahn V, Kramer HH, Voit T, Bruster HT, Scrampical B, Scheid A. Horizontal transmission of HIV infection between two siblings [Letter]. Lancet. 1986;2:694.

43. Vidmar L, Poljak M, Tomazic J, Seme K, Klavs I. Transmission of HIV-1 by human bite [Letter]. Lancet. 1996;347:1762-3.

44. Khajotia RR, Lee E. Transmission of human immunodeficiency virus through saliva after a lip bite [Letter]. Arch Intern Med. 1997;157:1901.

45. Saviteer SM, White GC, Cohen MS, Jason J. HTLV-III exposure during cardiopulmonary resuscitation [Letter]. N Engl J Med. 1985;313:1606-7.

46. Prevention of acquired immune deficiency syndrome (AIDS): report of interagency recommendations. MMWR Morb Mortal Wkly Rep. 1983;32:101-3.

47. Groopman JE, Salahuddin SZ, Sarngadharan MG, Markham PD, Gonda M, Sliski A, et al. HTLV-III in saliva of people with AIDS-related complex and healthy homosexual men at risk for AIDS. Science. 1984;226:447-9.

48. Ho DD, Byington RE, Schooley RT, Flynn T, Rota TR, Hirsch MS. Infrequency of isolation of the HTLV-III virus from saliva in AIDS [Letter]. N Engl J Med. 1985;313:1606.

49. Drummond JA. Seronegative 18 months after being bitten by a patient with AIDS [Letter]. JAMA. 1986;256:2342-3.

50. Transmission of HIV possibly associated with exposure of mucous membrane to contaminated blood. MMWR Morb Mortal Wkly Rep. 1997;46:620-3.

51. Sun D, Bennett RB, Archibald DW. Risk of acquiring AIDS from salivary exchange through cardiopulmonary resuscitation courses and mouth-to-mouth resuscitation. Semin Dermatol. 1995;14:205-11.

52. Mosley JW, Edwards VM, Casey G, Redeker AG, White E. Hepatitis B virus infection in dentists. N Engl J Med. 1975;293:729-34.

53. Smith JL, Maynard JE, Berquist KR, Doto IL, Webster HM, Sheller MJ. From the Center for Disease Control: comparative risk of hepatitis B among physicians and dentists. J Infect Dis. 1976;133:705-6.

54. Thomas DL, Gruninger SE, Siew C, Joy ED, Quinn TC. Occupational risk of hepatitis C infections among general dentists and oral surgeons in North America. Am J Med. 1996;100:41-5.

55. Villarejos VM, Visona KA, Gutierrez A, Rodriguez A. Role of saliva, urine and feces in the transmission of type B hepatitis. N Engl J Med. 1974;291:1375-8.

56. Powell E, Duke M, Cooksley WG. Hepatitis B transmission within families: potential importance of saliva as a vehicle of spread. Aust N Z J Med. 1985;15:717-20.

57. Scott RM, Snitbhan R, Bancroft WH, Alter HJ, Tingpalapong M. Experimental transmission of hepatitis B virus by semen and saliva. J Infect Dis. 1980;142:67-71.

58. Perrillo RP, Gelb L, Campbell C, Wellinghoff W, Ellis FR, Overby L, et al. Hepatitis B e antigen, DNA polymerase activity, and infection of household contacts with hepatitis B virus. Gastroenterology. 1979;76:1319-25.

59. MacQuarrie MB, Forghani B, Wolochow DA. Hepatitis B transmitted by a human bite. JAMA. 1974;230:723-4.

60. Cancio-Bello TP, de Medina M, Shorey J, Valledor MD, Schiff ER. An institutional outbreak of hepatitis B related to a human biting carrier. J Infect Dis. 1982;146:652-6.

61. Morris IM, Cattie DS, Smits BJ. Endoscopy and transmission of hepatitis B [Letter]. Lancet. 1975;2:1152.

62. McDonald GB, Silverstein FE. Can gastrointestinal endoscopy transmit hepatitis B to patients? Gastrointest Endosc. 1976;22:168-70.

63. Lack of transmission of hepatitis B to humans after oral exposure to hepatitis B surface antigen-positive saliva. MMWR Morb Mortal Wkly Rep. 1978;27:247-8.

64. Glaser JB, Nadler JP. Hepatitis B virus in a cardiopulmonary resuscitation training course. Risk of transmission from a surface antigen-positive participant. Arch Intern Med. 1985;145:1653-5.

65. Alter HJ, Purcell RH, Gerin JL, London WT, Kaplan PM, McAuliffe VJ, et al. Transmission of hepatitis B to chimpanzees by hepatitis B surface antigen-positive saliva and semen. Infect Immun. 1977;16:928-33.

66. Protection against viral hepatitis. Recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR Morb Mortal Wkly Rep. 1990;39(RR-2):17-22.

67. Hollinger FB, Lin HJ. Community-acquired hepatitis C virus infection [Editorial]. Gastroenterology. 1992;102(4 pt 1):1426-9.

68. Liou TC, Chang TT, Young KC, Lin XZ, Lin CY, Wu HL. Detection of HCV RNA in saliva, urine, seminal fluid, and ascites. J Med Virol. 1992;37:197-202.

69. Nakano I, Imoto M, Fukuda Y, Koyama Y, Urano F, Kanda N, et al. Hepatitis C virus RNA in urine, saliva, and sweat [Letter]. Am J Gastroenterol. 1992;87:1522.

70. Numata N, Ohori H, Hayakawa Y, Saitoh Y, Tsunoda A, Kanno A. Demonstration of hepatitis C virus genome in saliva and urine of patients with type C hepatitis: usefulness of the single round polymerase chain reaction method for detection of the HCV genome. J Med Virol. 1993;41:120-8.

71. Abe K, Inchauspe G. Transmission of hepatitis C by saliva [Letter]. Lancet. 1991;337:248.

72. Takamatsu K, Koyanagi Y, Okita K, Yamamoto N. Hepatitis C virus RNA in saliva [Letter]. Lancet. 1990;336:1515.

73. Young KC, Chang TT, Liou TC, Wu HL. Detection of hepatitis C virus RNA in peripheral blood mononuclear cells and in saliva. J Med Virol. 1993;41:55-60.

74. Couzigou P, Richard L, Dum