Risk and Management of Blood-Borne Infections in Health Care Workers
Elise M. Beltrami,1,* Ian T. Williams,2 Craig N. Shapiro,2 and Mary E. Chamberland1
http://cmr.asm.org/cgi/content/full/13/3/385#SEC9
HIV Infections Branch, Hospital Infections Program,1 and Hepatitis Branch, Division of Viral and Rickettsial Diseases,2 National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department of Health and Human Services, Atlanta, Georgia

SUMMARY
Exposure to blood-borne pathogens poses a serious risk to health care workers (HCWs). We review the risk and management of human immunodeficiency virus (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) infections in HCWs and also discuss current methods for preventing exposures and recommendations for postexposure prophylaxis. In the health care setting, blood-borne pathogen transmission occurs predominantly by percutaneous or mucosal exposure of workers to the blood or body fluids of infected patients. Prospective studies of HCWs have estimated that the average risk for HIV transmission after a percutaneous exposure is approximately 0.3%, the risk of HBV transmission is 6 to 30%, and the risk of HCV transmission is approximately 1.8%. To minimize the risk of blood-borne pathogen transmission from HCWs to patients, all HCWs should adhere to standard precautions, including the appropriate use of hand washing, protective barriers, and care in the use and disposal of needles and other sharp instruments. Employers should have in place a system that includes written protocols for prompt reporting, evaluation, counseling, treatment, and follow-up of occupational exposures that may place a worker at risk of blood-borne pathogen infection. A sustained commitment to the occupational health of all HCWs will ensure maximum protection for HCWs and patients and the availability of optimal medical care for all who need it.

INTRODUCTION
Exposure to blood-borne pathogens poses a serious risk to health care workers (HCWs). Transmission of at least 20 different pathogens by needlestick and sharps injuries has been reported (79). Despite improved methods of preventing exposure, occupational exposures will continue to occur.
Assessment of the risk of blood-borne pathogen transmission in the health care setting requires information derived from various sources, including surveillance data, studies of the frequency and preventability of blood contacts, seroprevalence studies among patients and HCWs, and prospective studies that assess the risk of seroconversion after an exposure to infected blood. Factors influencing the risk to an individual HCW over a lifetime career include the number and types of blood contact experienced by the worker, the prevalence of blood-borne pathogen infection among patients treated by the worker, and the risk of transmission of infection after a single blood contact.
In this article, we review the risk and management of the three blood-borne viruses most commonly involved in occupational transmission: human immunodeficiency virus (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV). We also will discuss current methods of preventing exposure, including standard precautions and the use of safety devices in the health care setting, as well as recommendations for postexposure prophylaxis.

TRANSMISSION OF BLOOD-BORNE PATHOGENS IN THE HEALTH CARE SETTING
Modes of Blood-Borne Pathogen Transmission
In the health care setting, blood-borne pathogen transmission occurs predominantly by percutaneous or mucosal exposure of workers to the blood or body fluids of infected patients. Occupational exposures that may result in HIV, HBV, or HCV transmission include needlestick and other sharps injuries; direct inoculation of virus into cutaneous scratches, skin lesions, abrasions, or burns; and inoculation of virus onto mucosal surfaces of the eyes, nose, or mouth through accidental splashes. HIV, HBV, and HCV do not spontaneously penetrate intact skin, and airborne transmission of these viruses does not occur.
Epidemiology of Blood Contact
To understand the nature, frequency, and prevention of percutaneous injuries and mucocutaneous blood contacts among HCWs, prospective observational studies have been performed in different patient care settings (Table 1). The percentage of procedures with at least one blood contact of any type ranged from 3% of procedures performed by invasive radiology personnel in a study in Dallas, Tex. (130), to 50% of procedures performed by surgeons in a study in Milwaukee, Wisc. (224). The percentage of procedures with at least one injury caused by a sharp instrument also varied widely, from 0.1 to 15%. These differences may be related to variations in study methods, procedures observed, and precautions used by the workers performing the procedures.

TABLE 1. Prospective observational studies of blood contact among HCWs
Several of these studies assessed specific risk factors for injury or exposure. For example, of the 99 percutaneous injuries observed by Tokars et al. during 1,382 operations in five different surgical specialties (general, orthopedic, gynecologic, trauma, and cardiac), most (73%) were related to suturing (256). Rates were highest (10%) during gynecologic surgeries (256). Panlilio et al. found in their study of blood contacts during surgery that risk factors for blood contacts by surgeons included performing an emergency procedure, patient blood loss greater than 250 ml, and surgery duration greater than 1 h (208). In their study of dental procedures, Cleveland et al. found that most percutaneous injuries sustained by dental residents occurred extraorally and were associated with denture impression procedures (77).
Retrospective studies and surveys have also shown high rates of blood contact among HCWs in different patient care settings. Tokars et al. found that among 3,420 participants at the American Academy of Orthopaedic Surgeons annual meeting, 87.4% of surgeons surveyed reported a blood-skin contact and 39.2% reported a percutaneous blood contact in the previous month (258). In a retrospective survey by O'Briain in 1991 (202), 56% of 36 resident and staff pathologists reported that they had sustained a cut or needlestick injury in the preceding year. In this study, pathologists reported 72 injuries, corresponding to a rate of one injury for every 37 autopsies performed and one injury for every 2,629 surgical specimens handled (202). An anonymous national survey of certified nurse-midwives by Willy et al. found that 74% had soiled their hands with blood, 51% had splashed blood or amniotic fluid in their faces, and 24% had sustained one or more needlestick injuries in the preceding 6 months (281). Among 550 medical students and residents in Los Angeles, Calif., who were surveyed anonymously by O'Neill et al., 71% reported exposures to patients' blood and body fluids during the preceding year (204). In a recent study of third- and fourth-year medical students in San Francisco, Calif., by Osborn et al., 12% reported an exposure to infectious body substances over the 7-year study period, from 1990 to 1996 (205). There is evidence among some groups of HCWs, such as dentists, that rates of exposure are decreasing over time, temporally associated with increased awareness and compliance with the practice of standard precautions (76).

DETECTION AND DIAGNOSIS OF BLOOD-BORNE PATHOGEN INFECTIONS
An understanding of the detection and diagnosis of HIV, HBV, and HCV infection is vital for the appropriate management and care of HCWs exposed to or infected with bloodborne viruses.
Detection and Diagnosis of HIV Infection
After initial primary infection with HIV, there is a window period prior to the development of detectable antibody. In persons with known exposure dates, the estimated median time from initial infection to the development of detectable antibody is 2.4 months; 95% of individuals develop antibodies within 6 months of infection (34). Among HCWs with a documented seroconversion to HIV, 5% tested negative for HIV antibodies at >6 months after their occupational exposure but were seropositive within 12 months (73). The two antibody tests commonly used to detect HIV are the enzyme immunoassay (EIA) and the Western blot. An HIV test result is reported as negative when the EIA result is negative. The result is reported as positive when the EIA result is repeatedly reactive and when the result of a more specific, supplemental confirmatory test, such as the Western blot, is also positive. Once an individual develops an antibody response, it usually remains detectable for life. HIV infection for longer than 6 months without detectable antibody is uncommon (73, 226).
Direct virus assays (e.g., PCR for HIV RNA) are sensitive methods for the detection of HIV infection. However, problems with laboratory contamination, false-positive rates, and increased costs limit their routine use. While PCR for HIV RNA is approved for use in established HIV infection, its reliability in detecting very early infection has not been determined (34). At present, the false-positive and false-negative rates of PCR are too high to warrant a broader role for it in routine postexposure management (207).
Detection and Diagnosis of HBV Infection
The incubation period for acute hepatitis B ranges from 45 to 160 days, with an average of 120 days. Exposure to HBV can lead to an acute infection which may result in a chronic infection. Acute hepatitis B resembles other forms of viral hepatitis and cannot be distinguished based on history, physical examination, or serum biochemical tests.
The diagnosis of acute HBV infection is confirmed by the demonstration in serum of hepatitis B surface antigen (HBsAg), which appears well before onset of symptoms and before development of antibody to hepatitis B core antigen (anti-HBc), and immunoglobulin M (IgM) antibody to HBc, which appear at approximately the same time as symptoms (143). The presence of IgM anti-HBc indicates recent HBV infection, usually within the preceding 4 to 6 months. The presence of hepatitis B e antigen (HBeAg) in serum correlates with HBV replication, high titers of HBV, and infectivity. Persons who are positive for HBeAg typically have 108 to 109 HBV particles per ml of blood (243). In persons who resolve acute HBV infection, antibody to HBsAg (anti-HBs) develops and indicates immunity. The persistence of HBsAg for 6 months after the diagnosis of acute HBV is indicative of progression to chronic HBV infection.
HBV serologic markers in different stages of infection and convalescence are summarized in Table 2. Anti-HBc indicates prior infection and lasts indefinitely. In persons who respond to the hepatitis B vaccine, anti-HBs is the only antibody that is elicited. Persons with chronic infection who have mutations in the precore region of the HBV genome that prevent the expression of HBeAg but allow the expression of infectious virus have been described (40, 260). High titers of HBsAg can be observed in these persons even though they are HBeAg negative. The prevalence of these precore mutations in persons in the United States is unknown. The prevalence may be relatively high in certain parts of the world (41, 124, 171, 173, 197).
 

TABLE 3. Tests for HCV infectiona

HCV RNA can be detected in serum or plasma within 1 to 2 weeks of exposure to the virus and several weeks before onset of alanine aminotransferase (ALT) elevations or the appearance of anti-HCV (103). In patients with chronic HCV infection, HCV RNA levels may remain relatively stable or can fluctuate over 1,000,000-fold. Fluctuations in HCV RNA may or may not correlate with elevations in transaminase levels. Rarely, the detection of HCV RNA may be the only evidence of HCV infection (14).
PCR techniques to amplify reverse-transcribed cDNA are currently the most sensitive methods for detecting HCV RNA. Both qualitative (122) and quantitative (87, 229) methods can be used to detect HCV RNA. Quantitative assays are less sensitive than qualitative assays and should not be used as a primary test to confirm or exclude the diagnosis of HCV infection (212). Currently, testing for HCV RNA is available on a research basis and no tests have been approved by the U.S. Food and Drug Administration. Because of assay variability, results of HCV RNA testing should be interpreted cautiously.
There are at least six different genotypes and more than 90 subtypes of HCV (33). About 70% of HCV-infected persons in the United States are infected with genotype 1; subtype 1a predominates over subtype 1b. Several different nucleic acid detection methods are commercially available to group isolates of HCV based on genotypes and subtypes (172).

RISK OF OCCUPATIONAL TRANSMISSION OF HIV FROM PATIENTS TO WORKERS
Risk of HIV Infection Postexposure
Prospective studies of HCWs have estimated that the average risk for HIV transmission after a percutaneous exposure to HIV-infected blood is approximately 0.3% (95% confidence interval = 0.2 to 0.5%) (23) and that after a mucous membrane exposure it is 0.09% (95% confidence interval = 0.006 to 0.5%) (147). The risk after a cutaneous exposure is less but has not been well quantified since no HCW enrolled in a prospective study has seroconverted after an isolated skin exposure. There are insufficient data to quantify the risk of transmission after occupational exposure to potentially infectious tissues or fluids other than blood. However, in a study by Fahey et al., none of 559 participants reporting cutaneous exposures to blood, sputum, urine, feces, or other body substances from patients presumed infected with HIV acquired HIV infection (102). There is also no evidence of a risk for HIV transmission by the aerosol route. Transmission of HIV by aerosol would require the generation of aerosolized particles of blood, the presence of infective HIV in these aerosolized particles, and the deposition of a sufficient number of infective particles in the respiratory tract or on the mucous membranes of a susceptible host to cause infection. Biological or epidemiologic evidence that HIV can be transmitted by aerosols via the respiratory route currently does not exist (22). Although not specifically designed to assess the possibility of aerosol transmission of HIV, the 1991 seroprevalence survey of attendees of the annual meeting of the American Academy of Orthopaedic Surgeons addressed this concern indirectly (258). There were 1,201 study participants without nonoccupational risk factors who had participated in procedures on patients with HIV infection or AIDS and had never used a "space suit" or other device to prevent inhalation of aerosols. Since power instruments are used frequently in orthopedic procedures, many of these participants may have been exposed to blood or tissue aerosols produced by these instruments; all were HIV seronegative (258).
The risk of HIV transmission after a percutaneous exposure appears to be influenced by several factors. To assess possible risk factors, the Centers for Disease Control and Prevention (CDC), in collaboration with international public health authorities, conducted a retrospective case-control study using data reported to national surveillance systems in the United States, France, Italy, and the United Kingdom. Based on logistic regression analysis, factors associated with HIV transmission after percutaneous exposure included a deep injury, a device visibly contaminated with the source patient's blood, procedures involving a needle placed directly in the patient's vein or artery, and a source patient who died from AIDS within 60 days of the exposure (39). The findings of the case-control study suggest that the risk for HIV infection likely exceeds 0.3% for percutaneous injuries involving a larger volume of blood and/or higher titer of HIV in the blood. Several laboratory studies support these findings. In vitro models have shown that increasing needle size and penetration depth are associated with increased blood transfer volume (182), that hollow-bore needles transfer greater volumes of blood than solid suture needles, and that gloves reduce the amount of blood transferred (26). Studies also have shown that the level of infectious HIV present in the blood of most patients with symptomatic AIDS is significantly higher than the level present in patients with asymptomatic HIV infection (141). An additional finding of the case-control study was that postexposure use of zidovudine (ZDV) by HCWs was associated with a lower risk for HIV transmission (39). (This issue will be discussed in more detail in the section Postexposure Chemoprophylaxis for HIV [below]). It is also possible that host defense mechanisms influence the risk of HIV transmission. One study demonstrated an HIV-specific T-helper cellular immune response when peripheral blood mononuclear cells from a small number of HCWs exposed to HIV were stimulated in vitro by HIV. None of the HCWs seroconverted. One possible explanation for these observations is that host immune responses prevented establishment of HIV infection after exposure (75). Similar cytotoxic T-lymphocyte responses have been observed in other populations with repeated HIV exposure without resulting infection (70, 74, 160, 170, 225).
HIV Seroprevalence among Patients
In the United States, HIV seroprevalence rates vary widely by geographic area and patients' demographic characteristics. The CDC's Sentinel Hospital Surveillance System tested 195,829 anonymous patient blood samples at 20 hospitals in 15 cities between September 1989 and October 1991. The HIV seroprevalence at these institutions ranged from 0.2 to 14.2% and was highest among men aged 25 to 44 years and patients with infectious conditions (excluding symptomatic HIV infection) and drug-related conditions (153).
Similarly, seroprevalence data for unselected hospital admissions and for patients presenting to emergency departments, operating rooms, and obstetrical units have demonstrated considerable variation (Table 4). The lowest seroprevalence rates have been reported in rural and suburban areas: 0.15% among trauma patients in Wichita, Kans. (190), and 0.4% among elective surgery patients in suburban Baltimore, Md. (68). The highest seroprevalence rates have been reported in urban, inner-city populations: 5.2 to 6.0% among emergency department patients in inner-city Baltimore, Md. (157, 191), and 5.5% among non-obstetric hospitalized patients in Denver, Colo. (K. Krasinski, W. Borkowski, D. Bebenroth, and T. Moore, Letter, N. Engl. J. Med. 318:185, 1988).
 

TABLE 4. HIV seroprevalence in emergency, hospital, surgery, and obstetrics patients


In a CDC study conducted in six emergency departments in three urban and three suburban areas of New York, N.Y., Chicago, Ill., and Baltimore, Md., the overall rate of HIV infection ranged from about 4 to 9 per 100 patient visits (178). The study found that many patients' HIV infections were unrecognized at the time of initial presentation to the hospital. The percentage of patients whose HIV infection was unknown to hospital emergency department workers was about 70% in the three inner city hospitals and ranged from 40 to 90% in the three suburban hospitals.
Incidence of Occupationally Acquired HIV Infection
As of 30 June 1999, a total of 191 U.S. workers had been reported to the CDC's national surveillance system for occupationally acquired HIV infection (Table 5) (65). Fifty-five HCWs had known occupational HIV exposures, with a baseline negative HIV test and subsequent documented seroconversion. Fifty of these exposures were to HIV-infected blood, one was to visibly bloody fluid, one was to an unspecified fluid, and three were to concentrated virus in a laboratory. Of the 55 HCWs, 47 sustained percutaneous exposures, 5 had mucocutaneous exposures, 2 had both a percutaneous and a mucocutaneous exposure, and 1 had an unknown route of exposure. Twenty-five of these HCWs have developed AIDS.
 

TABLE 5. HCWs with documented and possible occupationally acquired HIV infection reported through June 1999 in the United Statesa


Of the 191 U.S. workers reported to the CDC's surveillance system, 136 have been reported as possible cases of occupationally acquired HIV infection. None of these HCWs reported behavioral or blood transfusion risk factors, and all reported occupational exposures to blood, body fluids, or laboratory specimens containing HIV. However, the time or source of infection was undocumented, usually because no baseline serum sample was available to establish seronegativity at the time of exposure.
The CDC's surveillance system likely does not reflect the full extent of occupationally acquired HIV infection because of underreporting of known infections or underrecognition of HIV infection. Studies of HCWs in hospital settings suggest that many percutaneous injuries are not reported (129, 177). Also, HCWs may not complete postexposure follow-up serologic testing (D. Cardo and the Health Care Worker Surveillance Study Group, Abstr. 6th Annu. Meet. Soc. Healthcare Epidemiol. Am., abstr. 67, 1996).
HIV Seroprevalence Surveys among HCWs
HIV seroprevalence surveys provide a way of indirectly assessing the risk of occupationally acquired HIV infection. The CDC has conducted two voluntary anonymous seroprevalence surveys of surgeons in different specialties. In 1992, a seroprevalence survey was done among general surgeons, obstetricians, gynecologists, and orthopedic surgeons practicing in moderate to high AIDS incidence areas. Of the 770 participating surgeons, one general surgeon, who reported nonoccupational risk factors for HIV infection on an anonymous questionnaire, was HIV positive (209). In 1991, a seroprevalence survey was done among surgeons attending the annual meeting of the American Academy of Orthopaedic Surgeons. Of the 3,420 participants, two surgeons, both of whom reported nonoccupational risk factors, were HIV positive (258). Other seroprevalence studies similarly have shown low rates of HIV seropositivity among HCWs without nonoccupational risk factors for HIV infection (Table 6) (20, 66, 71, 80, 82, 107, 117, 118, 123, 163, 215, 264; P. Ebbensen, M. Melbye, F. Scheutz, A. J. Bodner, and R. J. Bigger, Letter, JAMA 256:2199, 1986; C. Siew, S. E. Gruninger, and S. A. Hojvat, Letter, N. Engl. J. Med. 318:1400-1401, 1988).
 

TABLE 6. Published HIV seroprevalence in selected HCWs

One limitation of seroprevalence studies is that the extent of occupational and nonoccupational exposure to HIV among tested workers is usually unknown. Also, the rates may be underestimates if individuals deferred testing because they knew they were or suspected they might be HIV positive. Nonetheless, these seroprevalence surveys indicate that there was not a high rate of undetected HIV infection among the HCWs studied, many of whom had substantial opportunity for occupational exposures.

RISK OF OCCUPATIONAL TRANSMISSION OF HBV FROM PATIENTS TO WORKERS
Risk of HBV Infection Postexposure
The probability of HBV transmission after an occupational exposure is dependent upon the concentration of infectious virions in the implicated body fluid, the volume of infective material transferred, and the route of inoculation (e.g., percutaneous or mucosal).
HBV is present in high titers in blood and serous fluids, ranging from a few virions to 109 virions per ml (142). The virus is present in moderate titers in saliva, semen, and vaginal secretions (154). The titer in semen and saliva is generally 1,000 to 10,000 times lower than the corresponding titer in serum (44, 269). Other body fluids such as urine and feces contain very low levels of HBV unless contaminated with blood (91, 106, 149).
One of the most common modes of HBV transmission in the health care setting is an unintentional injury of an HCW from a needle contaminated with HBsAg-positive blood from an infected patient (5). The average volume of blood inoculated during a needlestick injury with a 22-gauge needle is approximately 1 µl (V. M. Napoli and J. E. McGowan, Letter, J. Infect. Dis. 155:828, 1987), a quantity sufficient to contain up to 100 infectious doses of HBV (243). The risk of transmission after a needlestick exposure to a nonimmune person is at least 30% if the source patient is HBeAg positive but is less than 6% if the patient is HBeAg negative (17, 120, 277). Blood from patients with HBsAg titers below the threshold of detection using routine serologic tests is rarely infectious (4). While overt percutaneous injuries are efficient modes of HBV transmission, other less-obvious exposures may also lead to occupationally acquired HBV infection. In a case series of HBV-infected HCWs, fewer than 10% recalled a specific percutaneous injury, while 29 to 38% recalled caring for an HBsAg-positive patient within 6 months prior to their onset of illness (35; A. K. R. Chaudhuri and E. A. C. Follet, Letter, Br. Med. J. 284:1408, 1982).
HBV Seroprevalence among Patients
The risk of acquiring HBV is related to the prevalence of HBV infection in the patient population with which the HCW works. Patients who are HBsAg positive, either from acute or chronic infection, are potential sources of infection. Patients who are