Abstract
General considerations
Clinically significant viral infections are increasingly being recognized in a variety of patients who are immunocompromised. The immunosuppressed patient may experience cytomegalovirus (CMV) infection, and this disease state carries with it considerable morbidity and mortality, ranging from febrile illness to extensive dissemination with organ manifestations such as pneumonia, encephalitis, hepatitis and colitis (Ljungman et al., 2002). Included in this broad category of immunocompromised patients are allograft recipients of solid organs or bone marrow and stem-cell transplants (Shelby & Shanley, 1987; Mazzulli et al., 1993), and trauma patients (Heininger et al., 2001). Similarly, burn patients are known to be severely immunocompromised (O'Mahony et al., 1985; for review, see Munster, 1996). Currently, there is general agreement that the functional capacity of T cells in these patients to perform their normal physiologic responses is impaired (Zedler et al., 1999). Burn patients are particularly susceptible to infection and sepsis (Shires, 1991) as a result of their inherent immunosuppression as well as the loss of the cutaneous barrier and the use of invasive catheters, endotracheal tubes and other devices. Pneumonia, wound infection and sepsis are responsible for greater than 50 % of the mortality of patients suffering severe burn injuries (Huang et al., 1992). Routine treatment of severely thermally injured patients includes allogeneic skin application for temporary wound coverage of excised burn wounds. Allogeneic skin procurement follows the guidelines of the American Association of Tissue Banks and involves either cryopreservation or glycerol preservation (Vloemans et al., 2002). Previous reports on trauma patients have correlated septic episodes with CMV infection (Heininger et al., 2001), yet data linking morbidity to CMV infection in burns remains limited. A recent case of a patient with a 60 % total body surface area burn, septic episodes and concomitant blood CMV DNA levels of about 10 000 copies per millilitre of plasma (Hamprecht et al., 2005), and a survey of CMV infection at German burn centres (Pfau et al., 2004), prompted us to review the pertinent literature. This review summarizes the available experimental and clinical data on CMV infection as it relates to burns.
Animal experiments
Two murine animal studies utilizing a murine CMV (MCMV) strain (Shelby & Shanley,1987; Kobayashi et al., 1999) examined the role of skin-allograft-associated CMV transmission by temporary skin grafting. In the study by Shelby & Stanley (1987), it was shown that transplanted donor-skin isografts from MCMV-infected animals were able to infect non-burned mice with a transmission rate between 50 and 60 %. In their report, these authors raised the question of the cellular source of the virus in the transplanted skin tissue, but were unable to definitively answer this question. Interestingly, in vitro experiments (Kobayashi et al., 1999) with skin from CMV-latent mice failed to show replicating virus in culture by plaque methods in C127-I cells (Smee et al., 1989); however, MCMV DNA and mRNA were retrieved from 70 % of these murine skin samples. In the study of Kobayashi et al. (1999), allogeneic skin from MCMV-infected mice was transplanted onto both burned and non-burned animals. The authors could not identify MCMV growth in the salivary gland in the non-burned and grafted mice, but a significant number of MCMV copies was detected in burned and grafted animals.
Several murine studies demonstrate that infection with CMV enhances susceptibility to secondary bacterial infection. In a series of experiments, Overall and his group (Hamilton et al., 1976; Hamilton & Overall, 1978; Bale et al., 1982) demonstrated that MCMV-infected mice were highly susceptible to infection by intestinal challenge with Escherichia coli, intravenous challenge with Pseudomonas aeruginosa, or combined infection with Pseudomonas aeruginosa, Staphylococcus aureus and Candida albicans. They concluded that MCMV infection decreases the inflammatory response by a diminished leukocyte migratory response, and as a result inhibits host resistance to bacterial infections. Mortality was significantly enhanced in the MCMV-infected animals compared to those animals infected with either the bacterial or the fungal pathogen alone.
The translocation of bacteria from the gastrointestinal tract to mesenteric lymph nodes and systemically is considered to be a possible source of infection in burns and other types of major stress (Tancrede & Andremont, 1985). In burned mice challenged with MCMV, a significantly higher level of bacterial translocation to mesenteric lymph nodes was found compared to that of either control thermally injured mice without MCMV inoculation or non-burned mice injected with MCMV alone (Erickson et al., 1990). In a subsequent experiment by Erickson et al. (1991), the authors reported a significant delay in the resolution of positive mesenteric lymph node cultures in thermally injured animals infected with MCMV compared to thermal insult or MCMV infection alone.
A recent mouse study evaluated the pathophysiology of the higher susceptibility to MCMV infection after thermal injury (Kobayashi et al., 2001). These authors examined the role of the burn-associated T-cell response after burn injury. In their experiments, they found that 85 % of thermally injured BALB/c mice intraperitoneally inoculated with 2·5x105 p.f.u. MCMV died within 9 days. Unburned mice with a comparable inoculum had a significantly reduced mortality rate of 20 %. In subsequent experiments, the same authors showed that normal splenic T cells from BALB/c mice could improve resistance to MCMV infection in immunodeficient mice and subsequently reduce mortality, while splenic T cells from thermally injured BALB/c mice were not able to improve resistance to MCMV infection.
Clinical experience
Reports concerning human cytomegalovirus (HCMV) infection in burn patients can be divided into two major categories of research papers and some case reports. The first group of research papers studies the epidemiology of CMV infection among patients with burn injuries. The second group of research papers addresses the question of HCMV transmission by transplanted allogeneic skin.
The first report of HCMV in a thermally injured patient was published in 1970 (Nash et al., 1970). Later, Seeman & Konigova (1976) reviewed a series of 74 patients who suffered greater than 20 % total body surface area third degree burns, and who required skin grafting, blood transfusion and antibiotic regimens. Complement-fixing techniques revealed 27 % of the patients to be HCMV negative and 73 % to be HCMV positive. At least three of the deaths were assigned to pneumonia, and were felt to have resulted from a disseminated form of HCMV infection. In a follow-up paper, Seeman et al. (1980) reported a mortality rate attributable to systemic HCMV infection of 4·6 %.
Antibody titres in the serum of burned children have been analysed in both a retrospective study and a prospective study (Linnemann & MacMillan, 1981). In the latter, HCMV infection was diagnosed by a complement-fixing technique, and viral recovery was obtained from both throat swabs and urine. In the retrospective study, a significant increase in HCMV titres was identified in 22 % of the patients. In the prospective study, HCMV infection was diagnosed by virus isolation in 29 % of the patients, and serologically in 25 % of the patients, with an overall infection rate of 33 %. The prospective study directly correlated HCMV infection with more severe burns, more skin grafts and subsequent higher numbers of blood transfusions. The authors suggested that the suppression of cell-mediated immunity by the burn injury was a possible source of reactivation. In the prospective study, at least, the presumed source of primary HCMV infection seemed to be blood transfusions, as discussed earlier by Adler (1983). Both Linnemann & MacMillan (1981) and Deepe et al. (1982) observed a rise in HCMV antibodies during periods of unexplained fever and lymphocytosis.
In two similar surveys of burned patients (Kealey et al., 1987; Bale et al., 1990), it was shown that 22·5 and 18·6 %, respectively, of seronegative patients seroconverted. Among patients who were HCMV antibody positive, 56 and 52 %, respectively, reactivated, as assessed by a rise in titre of at least fourfold. However, differentiating passively acquired HCMV antibodies from production by an immunologic response was not addressed. Reactivation was associated with the severity of the burn injury and the number of transfusions, as well as length of the hospital stay (Kealey et al., 1987). For seronegative patients, seroconversion significantly correlated with the number of units of blood and blood products transfused, as well as the length of stay. No correlation could be established with relationship to total body surface area burn (Kealey et al., 1987). Kealey and his group attributed seroconversion to transfusion and transplantation of seropositive blood products and allograft skin, respectively. They stated that HCMV infection did not appear to contribute to the development of morbidity or mortality in their patient population. In contrast, Kagan et al. (1985) demonstrated a higher incidence of septic episodes in HCMV-infected patients, as shown by serology. An association between the number of blood transfusions and the emergence of HCMV infections was identified; however, no correlation was found with respect to mortality.
Kealey et al. (1996) studied the incidence of HCMV transmission by skin allografts, and found that 22·7 % of seronegative patients seroconverted after allograft transplantation. Unfortunately, the preservation methodology for the allografts was not stated. Significantly different viral potentials have been reported for fresh, cryopreserved and glycerol-preserved allograft temporary wound coverage (Van Baare et al., 1994).
Destruction of the skin is another hallmark of burn injuries, and timely healing is essential in reducing morbidity and mortality. An interesting case report of three patients presented by Swanson & Feldman (1987) analysed HCMV infection and skin involvement. A shared clinical course of generalized non-healing wounds prompted microscopic evaluations, which revealed enlarged endothelial and periendothelial cells containing prominent intranuclear and intracytoplasmatic inclusions diagnostic for HCMV. The authors concluded that HCMV infection of the skin may be more common than suggested by the few available reports. They did however note that it remains unclear in which way HCMV infection of the skin will systemically affect the individual patient.