Cytomegalovirus Infection

Abstract

After completing this article, readers should be able to: Cytomegalovirus (CMV) is the most common congenital infection in the United States, and it is a leading cause of sensorineural hearing loss, mental retardation, and cerebral palsy. In addition, CMV is a common opportunistic pathogen among immunocompromised patients. In both acquired immune deficiency syndrome (AIDS) patients and transplant patients, CMV infections are associated with illness that requires expensive treatment, compromises vital organ systems, leads to disability and diminished quality of life, and can be life-threatening.Spread of CMV is associated with close interpersonal contact. Not surprisingly, the prevalence of CMV infection and the age at first infection vary according to living conditions, child rearing practices, and sexual behavior. The overall prevalence of infection is greater and CMV is acquired earlier in developing countries and in the lower socioeconomic strata of developed countries. Virus is present in saliva, tears, semen, urine, cervical secretions, and blood for months to years after initial infection, and CMV is present in milk from seropositive mothers. CMV is transmitted via direct contact with body fluids (eg, breastfeeding), intimate contact, care of young children, blood transfusion, and organ transplant. Where breastfeeding is widely practiced and mothers are seropositive, the majority of infants acquire CMV during the first year of life. Young children who have CMV infection shed virus for years. Preschool-age children often acquire CMV from each other, especially in child care centers; they also are an important source of infection for adult caregivers. CMV also is spread by sexual contact. Nosocomial transmission of CMV, presumably by hand contact, has been reported in the hospital, but it appears to be uncommon. Rates of CMV infection are high among children and workers in child care centers. Although direct contact between children shedding CMV and susceptible classmates and caregivers is the likely mode of spread, CMV has been recovered from hands of workers and fomites in this setting, making transfer of virus to susceptible individuals by these routes possible.Congenital CMV infection occurs in approximately 0.5% to 1.5% of United States births. Rates of congenital CMV infection are highest in populations in which the prevalence of infection among women of childbearing age is high. In fact, rates of congenital CMV infection of approximately 2% have been reported in studies of African populations in which there is universal acquisition of CMV during childhood. It now is well known that CMV can be transmitted to the fetus even when maternal infection occurred long before conception. A large cohort study of low-income, predominantly African-Americans in Birmingham, Alabama, documented congenital CMV infection in approximately 1% of offspring of women who were seropositive at the time of birth of the previous child, an average of 2.5 years earlier. In middle-income groups, congenital CMV infection among offspring of immune mothers is much less frequent (∼0.2%).It is not clear why congenital CMV infection occurs in offspring of immune mothers; these infections could be the result of reactivation of latent virus during pregnancy, chronic infection, or reinfection with a new strain. Inflammation due to other infections also could facilitate CMV reactivation or reinfection in immune mothers and possibly account for the greater frequency of these infections in the low-income population. When primary maternal infection occurs during pregnancy, virus is transmitted to the fetus in about 35% of cases. The risk for transmission of CMV to the fetus does not appear to vary significantly with gestational age at time of maternal infection. The most important sources of CMV for primary maternal infections are sexual contacts and contact with young children. Rates of congenital CMV infection are three to seven times greater among babies born to adolescent mothers in the United States.Transmission of CMV from mother to infant plays a very important role in the epidemiology of CMV infection. Intrapartum transmission of virus from the maternal genital tract to the newborn and transmission via human milk occur far more commonly than congenital infection. Infants who acquire CMV shed virus in saliva and urine for years, making them an important source of virus for others in the community.CMV is a large, complex virus that has nearly 20 times the genetic material of the human immunodeficiency virus (HIV), with DNA sequences encoding more than 100 proteins. With human CMV infection, occurrence of disease is an aberration, resulting from infection in the fetus or a host who has impaired T-cell immunity. Important features of human CMV infection that relate to the pathogenesis are the ability of the virus to destroy host cells (lytic infection), to infect a wide range of cells and tissues, to evade and interfere with host defense mechanisms, and to persist indefinitely in the host. During productive infection (with release of progeny virus), the infected cell is destroyed.Histopathologic studies have found CMV in many different cell types and in many organs, including salivary glands, kidney, pancreas, adrenals, lung, liver, eye, ear, placenta, gastrointestinal tract, heart, ovaries, skin, blood vessels, and brain. In the central nervous system (CNS), CMV can infect different cell types, including neurons, glia, ependyma, choroid plexus, meninges, and vascular endothelium. Impaired organ function probably results from a combination of lytic infection of cells and vascular compromise due to infection of vascular endothelial cells. The ability of CMV to infect white blood cells and vascular endothelial cells probably facilitates dissemination of the virus within the host. A small proportion of circulating monocytes in seropositive persons harbor latent CMV; when these monocytes are activated and differentiate into tissue macrophages, CMV is reactivated from latent to productive infection, with release of infectious virus into surrounding tissue. The ability of CMV to maintain active infection in the normal host over long periods of time without elimination by the immune system may be due to multiple, complex strategies for interfering with host immune responses. These include inhibition of host cytotoxic T-lymphocyte responses by interference with the processing and presentation of viral antigens by major histocompatibility complex (MHC) class I molecules, interference with helper T-cell responses by degradation of MHC class II molecules, and inhibition of killing by natural killer cells. In addition, CMV encodes proteins that could have the ability to impede viral host defenses by other mechanisms, such as interference with apoptosis and complement-mediated lysis or sequestration of chemokines (host proteins involved in leukocyte activation, trafficking, and cell adhesion). Human CMV also encodes chemokines with the ability to attract neutrophils and monocytes, potentially facilitating the dissemination of virus. The presence of so many genes that appear designed for interaction with host cells and defense mechanisms suggests that CMV infection achieves a delicate balance, with persistence of virus and lack of disease being the expected result. When this balance is disturbed by inadequate host immune response or perhaps a viral strain with increased virulence, viral replication escapes control and spreads unchecked, resulting in disease. In immunologically normal children and adults, almost all acquired CMV infections are clinically inapparent. When symptomatic, acquired CMV infection presents as mononucleosis. It has been estimated that CMV accounts for 50% of cases of heterophile-negative mononucleosis. Mononucleosis due to CMV is indistinguishable clinically from the more common variety due to Epstein-Barr virus, although exudative pharyngitis and splenomegaly are observed less consistently with CMV. Fever and malaise often persist for more than 2 weeks. The expected outcome of CMV mononucleosis is complete recovery. Rarely, primary CMV infection in normal hosts, whether or not manifested as mononucleosis, has been associated with a variety of complications, including peripheral neuropathies, hemolytic anemia, thrombocytopenia, pneumonia, retinitis, gastrointestinal ulceration, and encephalitis.More than 90% of children who have congenital CMV infection exhibit no clinical evidence of disease as newborns. In newborns who have symptomatic congenital CMV infection, signs involving the skin, CNS, and hepatobiliary system predominate (Table 1). In addition to the findings listed in Table 2T1, approximately 40% of newborns are small for gestational age, 35% are preterm, and 25% of males have inguinal hernias. Laboratory abnormalities also reflect involvement of hepatobiliary, hematopoietic, and central nervous systems. Among newborns who have symptomatic congenital CMV infection, the severity of illness varies widely, ranging from one or two abnormalities noted on examination of an otherwise healthy newborn to multisystem disease with clear evidence of CNS damage. Mortality is 10% to 15% among symptomatic newborns. Among survivors, neonatal clinical abnormalities can be expected to resolve spontaneously within weeks, except for those involving the CNS and hearing.Congenital CMV infection is a leading cause of mental retardation and sensorineural hearing loss and an important cause of cerebral palsy and retinal damage. CNS and sensory impairments are seen in 50% to 90% of symptomatic newborns; these patients usually have multiple impairments. Among infants who are asymptomatic in the newborn period, the most common impairment is sensorineural hearing loss, which occurs in about 7% to 15%. Mental retardation and cerebral palsy are less frequent. In fact, it is not clear that children who have asymptomatic congenital CMV infection are at increased risk for these problems. Factors associated with increased risk of CNS sequelae include birth to a mother who has primary infection during pregnancy (although this is rarely identified outside the research setting) and the following newborn findings: abnormal neurologic examination findings, microcephaly, cranial computed tomography abnormalities, increased cerebrospinal fluid protein, and chorioretinitis. It is important to recognize that in many infants who have congenital CMV infection, the onset of hearing loss is after the newborn period or is progressive. The FigureF1shows audiometry results for an infant who had congenital CMV infection and passed the newborn nursery auditory brainstem response (ABR) screening examination, had a normal ABR threshold at 3.25 months of age, and progressed to bilateral severe-to-profound hearing loss during the first year of life.Perinatal CMV infection acquired during birth or from mother’s milk is not associated with newborn illness or CNS sequelae, except perhaps in very preterm newborns who have very low levels of passively acquired antibody at the time of infection. Transfusion-acquired CMV infection in newborns larger than 1,250 g who have serum antibody to CMV (born to seropositive mothers) also generally is without significant clinical consequences. However, among very low-birthweight preterm newborns born to seronegative mothers, transfusion-acquired CMV can cause newborn illness very similar to symptomatic congenital infection, with jaundice, hepatosplenomegaly, abnormal liver function, thrombocytopenia, and pulmonary problems. In addition, small preterm newborns who have transfusion-acquired CMV infection may develop CNS sequelae, including hearing loss.The clinical findings that are characteristic of congenital CMV infection also are seen in newborns who have congenital toxoplasmosis, rubella, and syphilis. Acute neonatal viral, bacterial, or systemic fungal infections occasionally produce similar findings. In addition, some inborn errors of metabolism can cause neonatal hepatitis, thrombocytopenia, hepatosplenomegaly, encephalopathy, and anemia. Although congenital CMV infection cannot be diagnosed reliably on the basis of signs and symptoms, the presence of microcephaly and hearing loss strongly suggests CMV. Because CMV is the leading cause of congenital infection in the United States, laboratory tests for the virus should be performed whenever clinical findings are compatible.CMV is a frequent and troublesome opportunistic pathogen for patients who have impaired T cell immunity, notably transplant and AIDS patients. Although a discussion of CMV infection in the immunocompromised host is well beyond the scope of this review, a few key points provide insight into the clinical problem and the biology of the virus. The majority of CMV infections in immunocompromised hosts are due to reactivation of the host’s CMV, and these often are clinically silent. However, when primary infection occurs in an immunocompromised host, it usually is associated with disease. The principal sources for primary CMV infection are blood products and the transplanted organ (when the donor is seropositive to CMV).The clinical manifestations of CMV infection, whether primary or reactivation, in both transplant and AIDS patients correlate with the degree of immunologic impairment. For example, among transplant patients, the most severe CMV disease is seen in allogeneic marrow recipients and in patients treated with aggressive immunosuppression. Among AIDS patients, the majority of CMV disease occurs in those whose CD4 counts are less than 100/mcL. Primary CMV infection in transplant patients frequently is heralded by fever and leukopenia; in addition, patients often develop rash, arthralgias, and elevated serum alanine aminotransferase levels. Serious complications of primary or reactivation infection in transplant patients include impaired function of the transplanted organ, pneumonitis, gastrointestinal ulceration, hepatitis, and opportunistic fungal infection. Among AIDS patients, the most frequent problems related to CMV are retinitis, esophagitis, and colitis; infrequent complications include encephalitis, peripheral neuropathy, hepatitis, and pneumonitis. Treatment of CMV disease in AIDS patients essentially is palliative, requiring induction therapy with ganciclovir, foscarnet, or cidofovir to control the disease, followed by maintenance therapy with one of these drugs. The maintenance therapy is continued indefinitely or until the patient’s immune function improves and stabilizes at a level that will allow discontinuation of anti-CMV treatment. Use of highly effective antiretroviral therapy has resulted in improved immunologic function in HIV-infected adults and children and has reduced the incidence of CMV disease dramatically. Over the past 20 years, the frequency of CMV disease among transplant patients has been reduced significantly by the use of antivirals for prophylaxis and treatment.Congenital CMV infection is diagnosed by detection of virus in the urine or saliva of a newborn. This can be accomplished by traditional virus culture methods, which can require up to 2 weeks to obtain a result. Rapid culture methods (“shell vial assay”) use centrifugation to enhance infectivity and monoclonal antibody to detect immediate early antigens in infected tissue culture cells and can give results in 24 hours. Polymerase chain reaction (PCR), with its ability to detect very small amounts of CMV DNA, is used by some laboratories, but it is not clear that it offers any advantage over culture-based methods because newborns who have congenital CMV infection shed large amounts of virus in both saliva and urine. Methods based on detection of immunoglobulin M (IgM) antibody should not be used to diagnose congenital CMV infection. They offer no advantages over virus culture and PCR, and they are less sensitive and more subject to false-positive results.Infants who acquire CMV during birth or from human milk also shed virus in body fluids starting several weeks after birth. Detection of CMV within the first 3 weeks of life is considered proof of congenital CMV infection. After 3 weeks of age, it is not certain that virus shedding is the result of prenatal or perinatal infection. Diagnosing congenital CMV infection in infants older than 3 weeks of age requires a combination of clinical judgment and laboratory testing for CMV and other etiologies in the differential diagnosis. The distinction between congenital and intrapartum or human milk-acquired CMV infection is significant because the latter are not associated with the CNS or auditory sequelae of congenital CMV infection.Diagnosis of CMV infections in immunocompromised patients is much more difficult than in the newborn who has possible congenital infection. The problem is not so much determining whether the patient has CMV infection as determining whether CMV is causing disease. Many immunocompromised patients shed CMV without suffering ill effects. The different approaches to diagnosis of CMV infection in immunocompromised patients are based on the setting and local laboratory preference. Although discussion of this subject is beyond the scope of this review, a few general points will provide a framework for further reading. Serology is of very limited use. Among HIV-infected patients, transplant patients, and organ donors, the presence of IgG antibody to CMV is a good marker for past infection and risk for CMV disease with immune impairment. It also identifies the problematic positive donor/negative recipient situation that carries a high risk for primary infection and CMV disease in the immediate posttransplant period. Detection of CMV in blood is a much better predictor of disease among immunocompromised patients than is shedding of virus in urine or other body fluids. The correlation between laboratory results and clinical disease has been strengthened by quantitation of virus in blood. Traditional virus culture is insensitive for detection of viremia and not suited for quantitation of virus in blood. Quantitative tests based on detection of white blood cells positive for CMV antigens by immunofluorescence (antigenemia) have been used widely, as has quantitative PCR.Three antiviral agents are approved for use in treating CMV infections in immunocompromised hosts: ganciclovir, foscarnet, and cidofovir. Ganciclovir also has been investigated for treatment of congenital CMV infection in clinical trials sponsored by the National Institutes of Allergy and Infectious Diseases. Although these agents are far from ideal in terms of therapeutic index, expense, ease of administration, toxicity, and pharmacodynamics, it is important to recognize that their use has reduced the burden of CMV disease substantially in immmunocompromised patients. Induction treatment and maintenance antiviral therapy for CMV retinitis in patients who have advanced AIDS have deferred the loss of vision and prolonged life. In bone marrow transplant patients, the occurrence of severe CMV disease has declined from around 25% of patients to less than 5% in most centers, largely due to the use of antivirals for prevention of CMV disease. The approach to antiviral treatment in immunocompromised patients can vary with the clinical setting (solid organ, marrow transplant, AIDS, immunodeficiency, or cancer chemotherapy), the type of CMV infection (primary or reactivation), clinical manifestations of CMV disease, laboratory evidence of CMV infection, and degree of immunosuppression. The treatment of CMV infections in immunocompromised patients should involve consultation with an expert.No antiviral agent is approved for treatment of congenital CMV infection. However, a phase III randomized trial of ganciclovir versus no treatment in newborns who had severe congenital CMV infection has been completed, and results (presented at international meetings) suggest a benefit from treatment. This multisite clinical trial enrolled newborns who had symptomatic congenital CMV infection and evidence of CNS involvement, such as microcephaly, intracranial calcifications, or neurologic abnormalities. Patients were randomized to 6 weeks of intravenous ganciclovir or no treatment, and the primary endpoint was hearing outcome. Results showed a highly significant association between ganciclovir treatment and lack of progression of hearing loss up to a mean follow-up of approximately 2 years. The toxicity observed most commonly in this clinical trial was neutropenia (63% of ganciclovir recipients and 20% of placebo recipients). The implications of these results for clinical practice are not yet clear. Questions remain about whether the benefit will be sustained over time. Treatment usually requires placement of a central line, with its associated procedural and infection risks. Some newborns in earlier trials had marked worsening of retinitis after ganciclovir treatment was stopped. In addition, ganciclovir has the potential for reproductive toxicity or carcinogenicity that might not be apparent until treated patients reach maturity. It also is important to recognize that this trial enrolled the most severely affected newborns. The risk:benefit ratio for these patients may be sufficiently low to justify treatment, but the risk:benefit ratio will be much greater for patients who are asymptomatic at birth or less severely affected than the clinical trial participants.Children who have congenital CMV infection are at risk for developmental delay, neurologic abnormalities, mental retardation, impaired vision, and hearing loss. Their follow-up should include routine assessments aimed at early identification of problems. A recommended schedule of follow-up evaluations is shown in Table 2. All newborns who have congenital CMV infection should undergo hearing testing using evoked response audiometry and eye examinations using indirect ophthalmoscopy. The primary care physician should include careful neurologic and developmental assessments in the plan for health supervision visits of these children. When abnormalities are found, it may be helpful to seek the advice of specialists in neurology or neuropsychology. Early detection of hearing loss is particularly important because of its impact on speech development. If abnormalities are detected, more frequent evaluations are recommended until the course and severity of the hearing changes can be determined. For children who have no hearing loss by 1 year of age, semiannual evaluations up to age 3 years, followed by annual evaluations up to school age are recommended.The primary sources of CMV infection for women of childbearing age are young children and sexual contacts. All preschool-age children should be considered potential sources of infection. Because transmission of virus appears to require contact with secretions, it is sensible to stress avoiding contact with body fluids from young children and careful handwashing whenever such contact could have occurred. The observation that CMV infection rates are not increased among hospital personnel who care for young children suggests that this approach can be effective. Prevention of CMV infection during intimate contact may be more difficult for women who are planning a pregnancy. However, it is reasonable to instruct women who are pregnant to use condoms and to avoid new sex partners. Although it is not certain that infections in pregnant women can be prevented by avoiding exposure, women who are or might become pregnant should be educated about avoiding exposure to CMV and the possible consequences of congenital infection. This education is most important for women who have occupational contact with young children or a history of sexually transmitted infections.A vaccine that could prevent maternal and congenital infection is clearly needed. Although none is on the immediate horizon, a phase II clinical trial of a recombinant subunit vaccine in young women is underway at the author’s institution, and a new, genetically engineered live virus vaccine recently entered phase I clinical trials.Secretion precautions and careful handwashing should be used in the clinical setting to prevent spread of CMV from patients to staff or other patients. Children known to have CMV infection should not be identified for special infection control procedures; rather, all children should be considered sources of CMV. Transfusion-acquired CMV infections can be prevented by limiting seronegative patients to blood products from seronegative donors or by removing leukocytes and platelets from whole blood by using special filters. It is customary to use one of these two approaches for blood products provided to preterm newborns and immunocompromised patients.The majority of CMV infections in immunocompromised hosts result from reactivation of endogenous virus. Although these infections cannot be prevented, prophylactic antiviral treatment and passive immunization with CMV immune globulin preparations have been used singly and in combination to prevent CMV disease after transplantation.