Cytomegalovirus (CMV) is a member of the herpes family of viruses. It is relatively common throughout the adult population (up to 50%), but occurs in a higher proportion of gay men (up to 95%) and drug users (64%). Like the other herpes viruses, it is transmitted sexually in semen and vaginal secretions, through blood or saliva, through organ transplants and from mother to baby before or during birth or through breastmilk.

In most people with fully functioning immune systems, initial infection with CMV may cause a mild flu-like illness, but afterwards it is kept dormant. If the immune system is damaged, as in people with advanced HIV disease, it can be reactivated. The average CD4 cell count of people at the time they develop their first episode of CMV is below 30 cells/mm³ (Kupperman 1993). Historically, over 20% of people with HIV have developed CMV disease within two years of their CD4 cell count falling below 100 cells/mm³ (Gallant 1992).

Cytomegalovirus in the age of highly active antiretroviral therapy

The advent of highly active antiretroviral therapy (HAART) regimens has had a significant impact on several aspects of CMV.

Firstly, fewer people with HIV have very low CD4 counts which put them at risk from CMV. The number of cases of new CMV disease has thus fallen to a much lower level in developed countries. A number of large studies conducted in the United States and France have found that the incidence of CMV retinitis dropped by between 73 and 83% from 1995 to late 1996 following the introduction of HAART (Palella 1998; Michaels 1999; Baril 2000b; Jalali 2000).

Secondly, people diagnosed with CMV retinitis are living much longer due to HAART. A study in London found that receiving a protease inhibitor is the strongest predictor of prolonged survival after a diagnosis of CMV retinitis (Walsh 1998).

HAART now plays a key role in preventing or treating CMV disease, regardless of what anti-CMV treatment is being used. In a large study comparing three CMV treatments, those who did not receive protease inhibitors were much more likely to progress, even in the group receiving the most effective anti-CMV therapy. Furthermore, successful antiretroviral therapy means that some individuals can stop taking CMV prophylaxis.

There have been reports of new types of eye problems, such as uveitis (inflammation of the iris, or entire eye) or inflammation within the back of the eye (vitritis), among people who had previously been treated for CMV retinitis within a year of starting HAART. This condition is caused by immune reactions against inactive CMV in the eye, rather than by CMV itself. Nevertheless, it can cause loss of sight. Since the CMV is usually dead, the optimal approach to treatment is unclear. It may be possible to treat these conditions with immune-suppressing steroids but some fear this could encourage CMV itself to reactivate. The risk of CMV reactivation appears to be determined by the degree of immune reconstitution and the presence of detectable CMV in the blood.

Cases have been reported between one and six months after initiation of successful HAART. Vitritis is relatively common among people with CMV retinitis who commence HAART. One study of 82 people with CMV retinitis who were treated with HAART found that six individuals developed vitritis (Jabs 2002).

Although HAART has been shown to reduce the likelihood of CMV re-activation over several years, a recent study has found that CMV-specific CD4 T-cell responses return to baseline levels after five years, peaking at about three years, in people who began HAART with a lowest-ever CD4 cell count below 50 cells/mm³. This was despite continued viral suppression and increased CD4 cell numbers and percentages. Additionally, increased numbers of CD8 T-cell responses may be an early indicator of recurrent CMV replication (French 2004).

CMV is not the only infection that can cause these eye problems. Other possible causes include Histoplasma, Leishmania, and possibly herpes zoster.

A recent study has shown that HIV-positive patients with low CD4 cell counts and detectable CMV are more likely to experience new AIDS-defining illnesses and to die, even in the era of HAART. This effect was not influenced by HIV viral load and CD4 cell count (Deayton 2004).

Symptoms

In immunosuppressed people CMV can cause disease in a number of different parts of the body. Up to 30% of people with AIDS develop CMV retinitis in the late stage of their disease, and an additional 5 to 10% develop CMV disease in other organs.

CMV retinitis is infection of the back of the eye, causing the retina to become swollen and inflamed. This means that the signals sent from the eye to the brain become incomplete or inaccurate, leading to blurring of vision or blind spots. If left untreated, the retina can become scarred, leaving permanent damage to the eyesight or no vision at all. A very early sign of CMV retinitis may be the appearance of 'floaters' in the vision, visual hallucinations of flashing lights, a misting of vision, or blurred or dark areas in the peripheral vision.

In some cases, people can have CMV retinitis that is on the verge of damaging their vision, but may not have experienced any kind of symptoms. For this reason, people with low CD4 cell counts are often advised to see an eye specialist for regular screenings.

Patients experiencing any symptoms should seek urgent examination by an ophthalmologist, as it is important to start treatment for CMV retinitis as early as possible. The doctor should look at the back of the eye every few days during treatment to make sure that the condition is improving.

Gastrointestinal CMV infection can lead to colitis (inflammation of the bowel), cholangitis (inflammation of the bile duct) and ulceration in the mouth, throat or around the rectum. These conditions can result in diarrhoea (often containing blood), fever, abdominal pain and weight loss.

CMV can often be found in the lungs of people with advanced AIDS but this only rarely seems to cause problems, usually in combination with Pneumocystis pneumonia (PCP). It can also sometimes affect the peripheral nerves, causing pain with numbness and weakness beginning in the legs which then spreads upwards, affecting bowel and urinary function.

CMV can also affect the brain and the central nervous system. Inflammation of the brain due to CMV is called CMV encephalopathy. Polyradiculopathy, also known as Cauda equina syndrome, has also been linked to CMV. This condition is usually associated with nerve pain due to a slipped disk. In HIV-infected individuals, this condition may be caused by CMV infection of the spinal cord, leading to symptoms of loss of feeling in the lower limbs, urine retention, and loss of reflexes. Polyradiculopathy is often fatal unless it is treated very promptly. Even after treatment there is a risk of relapse: preliminary research suggests that loss of hearing in one or both ears may be an early warning sign of relapse of polyradiculopathy. CMV may also lead to inflammation of several nerves in unrelated parts of the body.

Diagnosis

CMV retinitis is usually diagnosed by ophthalmological examination of the eye. Early lesions appear as small yellow-white patches with a grainy appearance, often with accompanying bleeding.

CMV disease can be hard to diagnose from symptoms alone, and often requires the virus to be detected in biopsy samples. CMV-infected cells can be identified under the microscope in samples of colon, oesophagus (throat) or lung tissue. Doctors may also try to grow or 'culture' CMV from blood, tissue or urine samples, but this can take several weeks.

For all these tests, a positive result does not necessarily establish that CMV is the cause of any specific symptoms. At present they are usually used to help confirm a diagnosis based on symptoms.

Treatment

The three standard treatments for CMV retinitis are ganciclovir (Cymevene), foscarnet (Foscavir) and cidofovir (Vistide). These drugs are produced in liquid formulations for intravenous infusion. Another injected treatment for CMV retinitis, fomivirsen (Vitravene) is approved as for use in people who have found standard treatments ineffective or unsuitable. However, due to the low incidence of CMV retinitis in Europe, the production and marketing of fomivirsen has ceased.

Intravenous ganciclovir and foscarnet are equally effective treatments for CMV retinitis and gastrointestinal CMV. Ganciclovir and foscarnet may be used in combination or alternated on a daily basis, as studies suggest that this improves the effectiveness of treatment. The combination of ganciclovir and foscarnet is also favoured by many clinicians as the optimal treatment for CMV affecting the central nervous system.

All the key anti-CMV drugs have been tested as intravitreal treatments, injected into the vitreal fluid of the eye, or as intraocular implants, which are surgically implanted into the eye and gradually release the drug. Although intravitreal ganciclovir alone has shown poor results, ganciclovir eye implants plus oral ganciclovir are as effective at delaying disease as ongoing intravenous treatment. However, ganciclovir implants are no longer available in Europe after the company holding the marketing authority decided to cease production in 2002. Fomivirsen implants were also approved in Europe and the United States but the manufacturer voluntarily withdrew the European marketing authorisation for fomivirsen in 2002. However the drug can be made available within Europe on a named-patient basis. Cidofovir is not used as a local therapy, after studies showed it to be toxic and ineffective.

Compared with intravenous treatments, ganciclovir implants are significantly better at delaying progression of retinitis in the treated eye, but intravenous treatment is more effective at delaying progression of retinitis in both eyes and preventing CMV disease elsewhere in the body.

The development of oral treatments for CMV has also slowed the demand for local treatment for CMV retinitis. Valganciclovir (Valcyte) is an effective oral pro-drug of ganciclovir, which has been approved in Europe and the United States (Martin 2002).

Maintenance therapy may involve implants or intravitreal injections of foscarnet, ganciclovir or cidofovir at low doses. However, long-term intravenous maintenance therapy usually requires patients to have a permanent catheter inserted so that they can perform the infusions at home. CMV can develop resistance to ganciclovir or foscarnet during long-term maintenance therapy and resistance is associated with progression of CMV disease. The International AIDS Society now recommends life-long maintenance therapy for individuals with retinitis who have not experienced immune recovery on HAART.

Oral ganciclovir may be taken as maintenance therapy although it is slightly less effective than intravenous ganciclovir in preventing CMV relapses. Oral ganciclovir causes fewer side-effects and may remove the need for a catheter and the accompanying inconvenience and risk of infections. However, for people with diarrhoea or malabsorption, intravenous ganciclovir may be preferred.

Given the success of antiretroviral therapy in restoring immune function, a person's history of anti-HIV treatment may affect the type of CMV treatment they receive. For example, a person with CMV who has not previously taken anti-HIV drugs could expect a strong immune response to treatment, and therefore they may not require long-term treatment provided by an implant (Martin 1999). An individualised approach to CMV treatment in people receiving antiretroviral therapy has been recommended by experts (Whitley 1999).

A recent study has suggested that the addition of the monoclonal anti-CMV antibody to the standard anti-viral regimen for CMV can reduce the incidence of death in newly-diagnosed CMV patients (Borucki 2004). However, the clinical usefulness of this therapy requires analysis in further studies.

Stopping maintenance treatment

United States Public Health Service guidelines issued in 1999 allow for CMV maintenance treatment or 'secondary prophylaxis' to be discontinued if a person has had a CD4 cell count above 100 to 150 cells/mm³ for three to six months. This decision should be made in consultation with an eye specialist and should consider the condition of the eyes, as well as the person's immune and virologic response to anti-HIV treatment.

Most clinics in the United Kingdom will recommend that CMV maintenance treatment should only stop when the CD4 cell count has remained above 100 cells/mm³ for at least six months. Some clinics also test for CMV in the blood or the strength of specific immune responses to CMV in order to decide when to stop maintenance treatment.

These recommendations are based on several studies which showed that individuals with sustained immune recovery remain free of active CMV retinitis for up to 90 weeks after stopping prophylaxis (MacDonald 1998; Tural 1998; Vrabec 1998; Whitcup 1999). During successful treatment with HAART, the risk of a recurrence of CMV retinitis is minimal (Tural 1998; Torriani 2000; Jouan 2001). Occasionally, primary episodes or relapses of CMV retinitis occur among people with CD4 counts over 200 cells/mm³ (Johnson 2001). If the CD4 cell count falls below 50 or 100 cells/mm³, prophylaxis should be reinstated.

Prophylaxis

Taking oral ganciclovir or valganciclovir to prevent a first attack of CMV disease among people with low CD4 cell counts is controversial. Both valganciclovir and oral ganciclovir have been licensed in the United States, but not the United Kingdom, as primary prophylaxis against CMV, despite the limited evidence of the efficacy of this strategy. One trial reported in early 1995 found that it approximately halved the risk of developing a first episode of CMV disease among people with CD4 cell counts below 50 cells/mm³, or people with CD4 cell counts below 100 cells/mm³ who had experienced an AIDS-defining illness. However, another trial found no significant reduction in CMV disease among people taking oral ganciclovir. These conflicting conclusions may be explained by differences in the study designs, including participants' CD4 cell counts, the duration of treatment and the ways in which CMV disease was diagnosed.

The effectiveness of oral ganciclovir may be predictable by tests for CMV DNA in the blood. One study found that oral ganciclovir was effective at preventing CMV retinitis in people with low levels of CMV DNA in the blood (less than 50,000 copies/ml), but not in people with higher levels. The researchers suggested that people with higher levels should be offered a pre-emptive course of intravenous or intravitreal CMV treatment, rather than a CMV prophylaxis regimen.

High doses of the anti-herpes drug aciclovir (Zovirax) have been reported to inhibit the growth of CMV. However, trials in people with HIV have found that high-dose oral aciclovir does not prevent CMV disease.

Valaciclovir (Valtrex) is a pro-drug of aciclovir that produces higher levels of aciclovir in the blood than can be achieved with oral aciclovir. A trial testing the effects of valaciclovir as compared with two doses of aciclovir for CMV prophylaxis, called ACTG 204, was stopped in early 1995 when there were more deaths in the valaciclovir arm than the aciclovir arms. When the trial was followed to its proper end, there was no statistically significant difference in survival between the groups, although valaciclovir decreased the occurrence of CMV retinitis by a third compared with either of two doses of aciclovir. However, in this trial valaciclovir was linked to bleeding in the kidney, a potentially fatal condition, and in the United States the drug is not recommended for immunosuppressed people.

Ongoing research aims to find ways to identify people who are at particularly high risk of developing CMV disease. As discussed below, researchers have develop ways of measuring the levels of CMV in their blood, in the hope that prophylaxis or pre-emptive treatment can be targeted to those in greatest need.

Testing for cytomegalovirus in the blood

Two tests for CMV can be used to identify people at high risk or developing CMV disease:

• CMV antigenaemia involves examination of white blood cells under the microscope, to look for the presence of CMV-specific proteins called pp65 or pp67.

• The polymerase chain reaction (PCR) is a chemical method that detects CMV DNA in a sample of blood, cerebrospinal fluid, urine or tissue biopsies.

People who have positive results on these tests may be advised to start 'pre-emptive' therapy with anti-CMV drugs (Finkelstein 1997; Bowen 1998; Chevret 1999). However, it is not clear whether a positive CMV test indicates the likelihood of disease. While two studies have associated detectable CMV with immune recovery and undetectable viral load on HAART, a third found a connection between CMV DNA detection and death (Bush 2000; Baldanti 2000; Wohl 2003).

The amount of CMV in the blood, rather than just its presence or absence, may help to predict the risk of disease. In an American study of people not taking anti-CMV drugs, only 40% of those who had CMV viraemia below 50,000 copies/ml developed CMV disease, compared with 100% of people with CMV viraemia greater than 150,000 copies/ml (Spector 1996b).

Cytomegalovirus as a co-factor for HIV progression

In 1989 doctors from the Royal Free Hospital in London published evidence from their cohort of people with haemophilia that suggested CMV may accelerate progression to AIDS. Additionally, it has been suggested that CMV may allow HIV to infect cells which do not have CD4 receptors.

However, the Royal Free team published updated figures in 1995, in which they failed to find a statistically significant link between CMV and increased risk of progression, while several other studies published in 1994 also failed to find any evidence that CMV accelerates HIV disease progression.

Although a more recent study found that CMV was associated with an increased risk of death, but not with lower CD4 count or higher viral load (Sabin 2000), the question of exactly how CMV may impact of HIV disease progression remains unclear.