William J. Britt
Human cytomegalovirus (CMV) is ubiquitous in the population, and individuals who become infected remain persistently infected for life, with intermittent shedding of infectious virus from mucosal surfaces. Although CMV rarely causes symptoms in normal individuals, it is an important cause of morbidity and sometimes death in immunocompromised hosts. CMV remains a well-recognized cause of disease in the newborn infant following intrauterine infection (congenital CMV) and the allograft recipients undergoing posttransplantation immunosuppression. CMV emerged as the most common opportunistic infection in HIV/AIDS patients prior to the advent of effective retroviral therapy. Invasive CMV infections can be observed in patients treated with immunosuppressive biologics such as anti-tumor necrosis factor (TNF) antibodies. In each of these clinical situations, the association of disease with CMV infection has been linked to high levels of virus replication and end organ disease, usually associated with virus dissemination. In contrast, there is likely another group of disease states associated with chronic effects of persistent CMV infection that reflects the robust inflammatory response induced by this virus. Such associations have included coronary artery disease, transplant vasculopathy and cardiac allograft loss, tubular sclerosis and renal allograft loss, exacerbations of inflammatory bowel disease, and possibly some cancers such as glioblastoma. In addition, there continues to be debate surrounding the role of CMV in immune senescence and the decrease in immune responsiveness observed in aging. Whether definitive evidence will eventually directly link CMV to these disease states is uncertain, but it is clear that understanding the complex biology of CMV infections, including the virus-mediated control of ensuing host responses to infection with this virus, will provide new insight into each of these diseases.
CMV is the largest of the human herpesviruses, with an estimated size of 190 nm. The 230 kb double-stranded DNA genome is about 50% larger than the herpes simplex virus genome and encodes over 200 open reading frames, which include 100 unique virion proteins and an unknown number of nonstructural proteins. Viral DNA replication takes place in the nucleus of the infected cell followed by virus assembly in both the nucleus and cytoplasm. The structure of the virus is typical of herpesviruses and includes a complex envelope composed of host cell–derived membrane studded with virion glycoproteins, an amorphous area between the envelope and the capsid called the tegument layer, and an icosahedral capsid that contains the virion DNA. The tegument layer is highly immunogenic and induces strong adaptive immune responses, including CMV specific CD8+ cytotoxic T lymphocytes that are thought to play a pivotal role in controlling CMV replication in the infected host. Likewise, the protein components of the viral envelope are also immunogenic and believed to induce protective antibody responses that have been correlated with virus neutralization. In vivo, CMV appears to replicate in nearly all tissue and cell types whereas in vitro productive virus replication (production of infectious progeny) occurs in primary cells derived from epithelial tissue and the dermis. Literature from the 1990s suggested that each strain of CMV isolated from epidemiologically unrelated individuals was genetically unique, a finding suggesting that an infinite number of distinct viruses existed in the human population. This observation has been validated with next-generation sequencing technologies, which have provided evidence that CMV exists as genetically diverse forms within an individual. This finding has argued that during replication, CMV DNA synthesis is fraught with error rates that are much higher than previous studies would predict and/or potential recombination events if permissive cells are infected with genetically diverse populations of viruses. Thus repeated exposures to CMV over time could result in an individual acquiring a library of CMVs as reinfection of previously infected individuals with new strains of CMV appears commonplace. These observations have led many investigators to argue that CMV must express an armamentarium of immune evasion functions that allow it to remain hidden from protective host immunity. This relationship between host and virus is best illustrated by the finding that over years a persistently infected individual can maintain a stable virus load, unwavering antiviral antibody responses, and in some individuals, up to 15% of a total peripheral blood CD8+ CTL activity dedicated to recognition of CMV infected cells, suggesting that a détente has been established between virus replication and host innate and adaptive antiviral immunity. Thus CMV efficiently persists in an infected host for a lifetime while inducing chronic immune activation. This latter characteristic of the biology of CMV infection has supported a linkage between CMV and many of the chronic diseases that have been associated with this ubiquitous virus.
CMV infections are acquired through several settings: (1) community exposure, (2) nosocomial transmission, and (3) intrauterine infection. Community acquisition occurs throughout life and is linked by exposure to CMV that is shed from mucosal surfaces such as saliva, genital secretions, and urine. Peaks in exposure occur during childhood and in adolescents and young adults, presumably in the latter cases secondary to sexual activity. Common routes of infection of the very young infant include perinatal exposure to infected genital secretions during birth and ingestion of CMV-containing breast milk. Breastfeeding is the most common route of CMV infection in early childhood. Ingestion of breast milk from seropositive women results in a rate of infection of about 60% in infants. Infection is most common during the first several months of breastfeeding, but the risk continues for the duration of breastfeeding. Infants infected through breast milk excrete virus in the saliva and urine for prolonged periods of time measured in months to years and thus serve as a reservoir of virus for spread to other infants, children, and adults. After this period of intense exposure to CMV during the first year of life, infection in the remainder of childhood and early teenage years depends on specific exposures such as enrollment in group childcare facilities and/or exposure to infected, similarly aged siblings. Up to 50% of young infants and children attending group care facilities can be excreting CMV, a source of virus that can result in infection of children enrolled in the facility and in some cases the adult workers in the facility. Furthermore, once infected at a group care facility, infants can then transmit virus to their parents and siblings, thus providing a mechanism for spread of CMV within the community. Throughout childhood and early adulthood, CMV is transmitted by exposure to saliva and urine. However, as noted above, in adolescence and early adulthood there is a spike in infection presumably associated with sexual exposure. CMV is considered a sexually transmitted infection, and a wealth of data has shown an increased rate of infection in the sexually active population as well as transmission in CMV-discordant couples. In summary, exposure to young children and sexual exposure represent the most consistent risk factors for acquisition of CMV infection.
Nosocomial infections with CMV are well described and are associated with exposure to blood products containing CMV and less commonly through allograft transplantation following transplantation of an organ from a CMV-infected donor. Prior to improvements in blood banking that limited the number of leukocytes in red cell transfusions and that more efficiently identified CMV infected donors, transmission of CMV by blood transfusion was not uncommon and closely related to the volume of blood that was transfused. Transfusion-acquired CMV infections often resulted in symptomatic illness, with laboratory findings including hepatitis and thrombocytopenia in children and adults. In newborn infants lacking antibodies to CMV secondary to being born to women without seroimmunity to CMV or in cases of extreme prematurity, severe and sometimes fatal infections could develop. Similarly, immunocompromised patients who received CMV-containing blood were also at risk for severe infection, regardless of their prior exposure to CMV. Methodologies that more efficiently deplete contaminating leukocytes and the use of blood products from CMV seronegative donors have greatly decreased the incidence of transfusion-associated CMV infections. Finally, CMV transmission through infected allografts is well described and infections arising from CMV transferred in the allograft are a major cause of morbidity in both the early and late period after transplantation. Severe infections and graft loss are more often associated with mismatches between the donor and recipient, as occurs if the donor has a history of CMV infection (donor, CMV positive) and the recipient has not been exposed to CMV (recipient, CMV negative; D+/R− mismatch). Even with effective antiviral therapy to modify CMV infections in the early posttransplant period, CMV infection remains linked to long-term graft dysfunction and graft loss, a particularly important problem in cardiac and lung transplant recipients.
Congenital CMV infection (present at birth) occurs following intrauterine transmission of CMV. Rates of congenital infection between 0.4% and 1.0% have been reported in the United States, with perhaps the best estimate being about 0.4% based on a large multicenter study. Rates as high 2% in some areas in Asia and Africa have also been described. Although the mechanism of transmission remains an active area of investigation, CMV is thought to be transferred to the developing fetus following hematogenous spread of CMV to the placenta, presumably followed by cell-free transfer of virus to the fetal blood system. The rate of transmission to the fetus is about 30% in women with primary infection during pregnancy; in utero infections also occur in previously immune women (nonprimary infection), albeit at a reduced rate that has been suggested to be on the order of 1–2%. This latter rate is an estimate because the number of previously immune women who experience active infection during pregnancy is not known. It is important to note that although the rate of transmission of CMV is more frequent following primary maternal infection, the absolute number of congenitally infected infants born to women with nonprimary infections in most populations outnumber those resulting from primary maternal infection by 3-4-fold. This is particularly true in Africa, South America, and Asia, where maternal seroimmunity to CMV often exceeds 95%. Interestingly, these populations also have the highest rates of congenital CMV infections. The source of nonprimary infection is also somewhat controversial. Older reports suggested that nonprimary infection followed reactivation (recurrence) of virus infection in seroimmune women, whereas more recent literature has demonstrated that reinfection by genetically distinct strains of CMV occurs in previously infected women and these newly acquired viruses can be transmitted to the developing fetus. In some studies, the reinfection rates are about 15–20%, with annualized rates as high as 25%. Thus immunity to CMV is far from protective, although it has been inferred from existing epidemiological data that it can modify the risk of transmission to the developing fetus.
The mechanism(s) of disease associated with CMV infections remains undefined for most clinical syndromes that follow CMV infection. Several reasons have contributed to the overall lack of understanding of the pathogenesis of CMV infections and include: (1) the asymptomatic nature of infections in almost all immunocompetent individuals; (2) the complexity of the underlying disease processes in immunocompromised hosts that often confound the assignment of specific manifestations of CMV infection; (3) the species-specific tropism of human CMV; and perhaps most importantly, (4) limitations inherent in observational studies in humans. The strict species specificity of most CMVs has been a major limitation in the development of animal models that closely recapitulate human CMV infections. However, models have been developed in nonhuman primates, guinea pigs, and rodents to address specific aspects of the biology of CMVs. Although CMV replicates in a limited number of cells types in vitro, CMV inclusions, antigens, and nucleic acids can be demonstrated in almost all organ systems and cell types in individuals with severe, disseminated infections. The virus does not exhibit strict cellular or organ system tropism in vivo. Hematogenous dissemination has been argued to be associated primarily with cell-associated virus, and significant levels of plasma virus are usually detected only in severely immunocompromised hosts with high total-blood viral loads. Virus and viral DNA can be recovered from neutrophils, monocytes, and endothelial cells present in peripheral blood. High levels of virus replication can result in end-organ disease, secondary to direct virus-mediated cellular damage. These manifestations of CMV infections are thought to result from uncontrolled virus replication and dissemination, secondary to deficits in innate and adaptive immune responses to CMV. In some cases, clinical disease has been observed in patients without significant levels of virus replication, a finding suggesting indirect mechanisms of disease such as immunopathologic responses to CMV. Such a mechanism was clearly operative in patients with immune recovery vitritis, a pathological T-lymphocyte–mediated response to CMV in HIV/AIDS patients with CMV retinitis that closely followed the reconstitution of their virus-specific T lymphocyte responses following active retroviral therapy. Likewise, the level of virus replication has not been closely correlated with several chronic diseases thought to be linked to CMV, an observation that is consistent with indirect mechanisms of disease such as immunopathologic responses. These mechanisms are better described in animal models of human CMV disease.
From early observations in patients with invasive CMV infections in allograft recipients it was apparent that immunosuppressive therapies that resulted in altered T lymphocyte function predisposed these patients to severe infections. These observations that were first described in the 1970s were confirmed in multiple studies over the following decade. Definitive evidence consistent with this mechanism was provided by a clinical study that demonstrated that in vitro expanded, CMV-specific cytotoxic T lymphocytes could limit invasive infection in hematopoietic cell transplant recipients. Invasive infections such as retinitis and colitis in HIV/AIDS patients with very low CD4+ T-lymphocyte counts also clearly demonstrated the importance of T lymphocyte responses and invasive CMV infections. Other studies in solid organ transplant (SOT) recipients have demonstrated that the passive transfer of immune globulins containing high titers of anti-CMV antibodies could provide some degree of protection from invasive disease, a finding that was consistent with the proposed role of antiviral antibodies in limiting CMV dissemination and disease in animal models of invasive CMV infections. The importance of innate immune responses such as natural killer (NK) cells and γδ T lymphocytes in limiting invasive infections have been well documented in representative animal models but definitive evidence for a key role in resistance to CMV infections in humans is limited. Lastly, effector molecules such as γ-interferon appear to a play an important role in controlling local CMV infections in animal models, but evidence of a similar role in humans has not been shown experimentally.
The control of acute CMV infection is clearly dependent on an effective adaptive immune response; however, even a vigorous T lymphocyte response is not sufficient to eliminate CMV from the infected host, as CMV persists for the lifetime of the host either as a low level chronic infection or as a latent infection with limited transcription from specific regions of its genome. The inability of the host to completely clear CMV remains incompletely understood, but the large array of immune evasion functions encoded by this virus likely contributes to the blunted innate and adaptive immune response. These functions include: (1) inhibition of apoptotic and necroptotic functions of infected cells; (2) inhibition of interferon regulated responses; (3) inhibition of NK cell activation; (4) downregulation of class I MHC expression, inhibition of class II MHC function; and (5) mechanisms to limit antibody recognition of envelope proteins such as carbohydrate masking of antibody recognition sites and extensive variation in amino acid sequences in virion envelop proteins. Although each of these functions by itself could potentially have only limited effects on virus clearance secondary to the redundancy of antiviral activities of the host immune system, when acting in concert they likely provide the virus an advantage that leads to its persistence. The importance of these evasion functions has been shown in animal models, and specific immune evasion functions have been shown to facilitate virus dissemination and persistence, reinfection with genetically similar viruses, as well as reinfections with new strains of virus in animals with existing immunity to CMV.
The clinical manifestations of CMV infection reflect the level of virus replication and the end-organ involvement. The manifestations of invasive CMV infections have been most commonly identified with syndromes that could be associated with a primary infection defined as an infection in an individual without existing immunity to the virus. Chronic CMV infections that have been associated with disease syndromes almost always have concurrent manifestations of the underlying causes of the primary disease, thus confounding the role of CMV in the primary disease process and its contribution to the clinical syndromes in these patients.
In the overwhelming majority of patients with acute CMV infections, there are no specific symptoms or clinical findings. In patients with symptomatic, acute CMV infection, clinical findings most commonly resemble a mononucleosis-like syndrome, with fatigue and occasionally cervical adenopathy. It has been reported that up to 20% of heterophile antibody negative mononucleosis could be attributed to CMV. Laboratory findings could include mild elevation of hepatic transaminases and decreased platelet counts.
The clinical presentation of CMV infection in immunocompromised hosts often reflects the magnitude of the immunodeficiency. Profoundly immunocompromised hosts such as hematopoietic stem cell transplantation (HSCT) recipients can present with disseminated infection and clinical manifestations reflecting disease in multiple organ systems, including liver, lung, gastrointestinal tract, and, less frequently, the CNS. Organ-threatening and life-threatening disease is not infrequent. In less immunocompromised patients such as most SOT recipients, CMV infection can present with fever, hematological abnormalities including leukopenia and thrombocytopenia, and mild hepatocellular dysfunction. In contrast to renal and liver SOT recipients, heart-lung and lung transplant recipients are at high risk for severe manifestations from CMV infection, presumably because the transplanted organ is a site of virus replication, disease, and life-threatening dysfunction. Prior to widespread use of antivirals for prophylaxis of allograft recipients, clinical disease usually developed between 30 and 60 day posttransplantation. More recently, prolonged antiviral prophylaxis has nearly eliminated CMV disease in the early posttransplant period in most SOTs, but late manifestations of CMV infection often become apparent after discontinuation of antiviral prophylaxis. These late manifestations are most worrisome in HSCT recipients, as they may signal deficits in graft function leading to invasive CMV infections. Finally, long-term graft function has been reported to be influenced by CMV infection. This has been most well studied in the renal allograft recipients and is thought by some investigators to represent a significant cause of chronic graft dysfunction and loss. Perhaps the most dramatic impact of CMV infection late in the posttransplant period can be seen in heart transplant recipients, where CMV is believed to play a major role in transplant vascular sclerosis, a vasculopathy of the coronary arteries in the allograft leading to loss of the transplanted heart.
Congenital infection with CMV can present with symptomatic infections (Table 282.1 ) in about 10% of infected newborns, whereas 90% of infected infants will have no clinical manifestations of infection in the newborn period and can be identified only by newborn screening programs. Severe multiorgan disease is infrequent and occurs in less than 5% of infants with congenital CMV infections. The clinical findings in infants with symptomatic congenital CMV infections can include hepatosplenomegaly, petechial rashes, jaundice, and microcephaly. These findings were utilized for decades in natural history studies to classify infants has having symptomatic or asymptomatic infections; however, more recently several authors have included intrauterine growth restriction as a finding of symptomatic congenital CMV infection. Laboratory findings are consistent with the clinical findings and include direct hyperbilirubinemia, elevation of hepatic transaminases, thrombocytopenia, and abnormal findings on cranial ultrasonography/computed tomography. If cerebrospinal fluid is obtained, there can be evidence of encephalitis, with elevation of mononuclear cell number and, in some cases, elevation of protein. A small number of symptomatically infected infants (<10%) will be found to have chorioretinitis. Finally, because hearing loss is the most common long-term sequelae associated with congenital CMV infection, the failure of an infant to pass a newborn hearing screening exam should alert caregivers to the possibility of congenital CMV infection. Hearing loss in the older infant and young child should also alert the clinician to the possibility of congenital CMV infection as about 50% of infants with hearing loss associated with congenital CMV infection will pass an initial hearing screening exam but develop hearing loss in later infancy and early childhood. Importantly, hearing loss can be progressive in infants with hearing loss secondary to congenital CMV infections. Lastly, the diagnosis of congenital CMV infection must be made within the first 2-3 wk of life, and congenital CMV infection cannot be assumed to be the cause of hearing loss in older infants without evidence of CMV infection in the newborn period.
Table 282.1
Findings in Infants With Symptomatic Congenital Cytomegalovirus Infection
| FINDINGS | % OF INFANTS |
|---|---|
| CLINICAL FINDINGS | |
| Prematurity (<37 wk) | 24 |
| Jaundice (direct bilirubin >2 mg/dL) | 42 |
| Petechiae | 54 |
| Hepatosplenomegaly | 19 |
| Purpura | 3 |
| Microcephaly | 35 |
| IUGR | 28 |
| 1 clinical finding | 41 |
| 2 clinical findings | 59 |
| LABORATORY FINDINGS | |
| Elevated ALT (>80 IU/mL) | 71 |
| Thrombocytopenia (<100,000 k/mm3 ) | 43 |
| Direct hyperbilirubinemia (>2 mg/dL) | 54 |
| Head CT abnormalities | 42 |
Findings in 70 infants with symptomatic congenital CMV infection identified during newborn screening program for infants with congenital CMV infection at the University of Alabama Hospitals over an approximate 20 yr interval.
CMV, cytomegalovirus; IUGR, in utero growth retardation; ALT, alanine aminotransferase.
An organized plan for follow-up is an important component of the clinical management of infants with congenital CMV infection. Because permanent sequelae are limited to disorders of the nervous system, long-term follow-up should include appropriate assessment of development and neuromuscular function in infected infants, with referral to specialized care if necessary. Hearing loss will develop in about 11% of infected infants, and in some infants hearing loss will progress during infancy. Thus audiologic testing and follow-up are mandatory in these patients. Other sequelae such as vision loss are infrequent, but vision testing and comprehensive eye examinations should be included in the care plan of infants with congenital CMV infection.
Perinatal infections can be acquired during birth or following ingestion of CMV-containing breast milk. In almost all cases perinatal infections have not been associated with any clinical manifestations of infection and perhaps more importantly, have not been associated with any long-term sequelae. In rare cases such as is seen in breast milk transmission of CMV to extremely premature infants or infants born to nonimmune women, perinatal infection can result in severe, disseminated infections associated with end-organ disease and death. These more severe infections are thought to develop in infants that lack transplacentally acquired antiviral antibodies, either secondary to extreme prematurity, or as the product of a mother lacking anti-CMV antibodies. However, definitive evidence supporting this explanation is lacking.
In the nonimmunocompromised individual, diagnosis of CMV infection requires evidence of a primary infection. Serological reactivity for CMV is lifelong following primary infection; therefore the presence of immunoglobulin G (IgG) antibody to CMV does not provide evidence of acute infection. In addition, IgM reactivity for CMV can be detected for prolonged periods after acute infection and cannot be used to reliably estimate the duration of infection. Furthermore, recovery of virus from body fluids such as saliva or urine does not in itself permit diagnosis of CMV infection, because persistently infected individuals can intermittently shed virus. In the immunocompromised host, CMV can frequently be recovered from patients in the absence of evidence of invasive CMV infection. Thus assignment of CMV as a cause of disease in this patient population must be made carefully, and other potential causes of symptoms and clinical findings in these patients must also be considered. Serological assays are of limited value in the transplant recipient secondary to impact of immunosuppression on antibody responses in the allograft recipient. Moreover, IgM antibodies can be produced following a nonprimary infection in these patients. Sequential viral load measurements by polymer chain reaction (PCR) in relevant body fluids such as blood and measurements of CMV DNA in biopsy tissue can be of great value in establishing CMV as a cause of disease in allograft recipients. Histopathological detection of characteristic owl eye inclusion–bearing cells is insensitive but can point to a diagnosis of invasive CMV infection. The addition of immunohistochemistry for the detection of CMV-encoded proteins and/or in situ hybridization for the detection of CMV nucleic acids has greatly improved histologic detection of CMV in tissue specimens.
The diagnosis of congenital CMV infections requires the recovery of replicating virus and/or viral nucleic acids within the first 2-3 wk of life. Sources of virus and viral nucleic acids include urine, saliva, and blood. Methods of detection include routine virus culture combined with immunofluorescence and PCR. Although quantification of virus in various specimens can suggest the likelihood of long-term sequelae such as hearing loss for a population of infected newborns, the predictive value for the individual patient remains limited. A considerable amount of effort has been devoted to identifying screening assays that would be suitable for populations of newborn infants. Initial interest centered on dried blood spots, because these samples are routinely collected as a component of newborn screening programs. Unfortunately, studies have indicated that the sensitivity of dried blood spots is too low to be considered useful for screening. In contrast, newborn screening using saliva has proven sensitive and specific and is now standard for newborn screening in some institutions. Identification of an infected infant by screening of saliva requires confirmation, preferably by assaying urine for the presence of CMV.
Early studies suggested that congenitally infected newborn infants could be identified by CMV-specific IgM reactivity and that elevated levels of CMV-specific IgM correlated with severity of disease. Subsequent studies have demonstrated that although this assay was of some value, the limited sensitivity of most assays employed to detect newborn IgM has also limited their clinical utility.
In nonimmunocompromised patients, demonstration of CMV-specific IgG seroconversion or the presence of CMV-specific IgM antibodies represents evidence of a newly acquired CMV infection. IgM anti-CMV antibody reactivity can persist for months depending on the sensitivity of the particular assay, thus limiting the use of IgM detection to precisely time the acquisition of CMV. The use of the IgG avidity assays in which CMV-specific binding antibodies are eluted with increasing concentrations of chaotropic agents such as urea can be used to estimate the duration of infection. This assay has been used almost exclusively in the management of CMV infections during pregnancy to aid in defining primary maternal infections. Detection of CMV in urine, saliva, blood, and tissue specimens obtained at biopsy can most reliably be accomplished by PCR-based methods, and because findings can be quantified, treatment responses can be monitored. However, conventional culture of CMV using human dermal fibroblasts often combined with immunofluorescence detection of CMV-encoded immediate early antigens also remains standard in many institutions. Routine histological stains allow detection of characteristic nuclear inclusions in tissue specimens (see above).
Treatment of immunocompromised hosts with invasive CMV disease has been shown to limit both the morbidity and the mortality in the patient with disseminated CMV infections with end-organ disease. This has been shown in allograft transplant recipients and patients with HIV/AIDS. Similarly, antiviral prophylaxis can limit the development of clinically important CMV disease in allograft recipients and is the standard of care in most transplant centers. Several agents are currently licensed for CMV infections, including ganciclovir, foscarnet, and cidofovir, and all have appreciable toxicity. Newer agents such as letermovir have been licensed for use in adults and it is expected that indications for this agent will extend into pediatrics. In some transplant centers, high-titer CMV immunoglobulins have been included as a component of prophylaxis. Early on, when the treatment of CMV infections with antiviral agents was in its infancy, treatment with CMV immunoglobulins was shown to alter the natural history of CMV infection in renal and liver allograft recipients. The effectiveness of antiviral agents when used as prophylaxis in the immediate posttransplant period has resulted in less frequent use of these biologics.
Treatment of congenitally infected infants with ganciclovir has been studied in several clinical trials, and a significant number of infected infants have been treated off-label with this agent because of severe CMV infections. Two studies conducted by the Collaborative Antiviral Study Group (CASG) sponsored by the NIH have suggested that 6 wk of intravenously administered ganciclovir or 6 mo of an oral preparation of ganciclovir could limit hearing loss and possibly improve developmental outcome of infected infants. Long-term outcomes of treated infants are not known; thus it is difficult to definitively interpret these studies. In addition, infants with severe perinatal CMV infection following breast milk ingestion with documented end-organ disease have been successfully treated with ganciclovir. Currently, there are no recommendations for the treatment of infants with congenital CMV infection, although the results from a larger study that will determine the efficacy of treatment in infants with asymptomatic congenital CMV infections may provide sufficient data to firmly establish treatment guidelines.
As was described in the preceding section, passive transfer of anti-CMV antibodies has been utilized to limit disease but not infection in allograft recipients. A similar approach has also been considered for prevention of intrauterine transmission of CMV and disease based on studies in animal models and limited observational data that suggested a role of antiviral antibodies in limiting disease following CMV infections in the perinatal period. An uncontrolled trial of human immune globulin reported in 2005 provided provocative evidence that passive transfer of anti-CMV antibodies to pregnant women undergoing primary CMV infection could limit transmission and disease. This study was seriously flawed in design, and findings from this trial were controversial. A second study utilizing the same immune globulin preparation failed to demonstrate that immune globulins provided protection from intrauterine transmission or disease. Thus it remains to be determined if passively transferred anti-CMV antibodies can modulate infection and disease following intrauterine exposure to CMV. A larger multicenter trial sponsored by the NIH (NICHD) has been terminated and results of this study should be available in the near future.
Active immunization for the prevention of congenital CMV infection (and in transplant recipients) has been a goal of biomedical research for over 3 decades. A number of different vaccine platforms have been explored, including replicating attenuated CMV as vaccines, protein-based vaccines, heterologous virus-vectored CMV vaccines, and DNA vaccines. In all cases, some level of immunity has been induced in volunteers. Larger scale trials have been carried out using replication competent, attenuated CMV vaccines and adjuvanted recombinant protein vaccines. Current approaches are directed toward development of an adequately attenuated replicating CMV that retains sufficient immunogenicity to induce protective responses. In contrast to current status of candidate attenuated CMV vaccines, considerable progress has been made in the testing of adjuvanted recombinant viral proteins. An adjuvanted recombinant glycoprotein B, a major protein component of the envelope and target of neutralizing antibodies, has been shown to induce virus neutralizing antibodies and CD4+ T lymphocyte proliferative responses. Moreover, this vaccine reduced virus acquisition by about 50% in a trial carried out in young women. However, closer examination of this vaccine trial revealed that protection was very short-lived and that the effectiveness of the vaccine was not convincingly demonstrated because of the small numbers of subjects in the trial, despite the statistical significance. A follow-up trial in adolescent women using the same vaccine preparation failed to show any statistically significant difference between vaccine and placebo recipients. Finally, a major question that will face all vaccine programs is whether existing immunity in seropositive women can be augmented to a level to prevent damaging infection in their offspring. The maternal population with existing immunity to CMV prior to childbearing age is responsible for the greatest number of congenitally infected infants in almost all regions of the world; thus merely recapitulating naturally acquired adaptive immunity to CMV with a vaccine may not be sufficient to prevent congenital CMV infection and/or limit disease.
Studies of the natural history of CMV have repeatedly demonstrated that transmission requires close, often direct contact with infected material such as secretions from the oral or genitourinary tract. Although limited data suggest that CMV can be transmitted on fomites, infectivity can persist for hours on surfaces such as toys. Limiting exposure to such secretions and attention to hygiene such as handwashing can drastically limit acquisition of CMV. Counseling has been shown to be very effective in the prevention of CMV infection in women of childbearing age. In fact, counseling programs have been shown to be more effective in limiting CMV infection during pregnancy than any vaccine that has been tested to date. Sexual transmission is an important route of infection, and CMV is considered to be a sexually transmitted infection. Limiting sexual transmission through education and counseling should be considered in sexually active individuals. Finally, the acquisition of CMV by hospital workers and other healthcare providers has been shown to be less than in age-matched individuals in the general public. Importantly, these studies were carried out prior to universal precautions that are in place in most hospitals today. Thus patient education with an emphasis on describing the sources of infectious virus in communities and attention to general hygiene could dramatically reduce CMV spread in the community, and particularly in women of childbearing age.