Design
Prospective cohort and retrospective casecontrol series.

Setting
University hospital pediatric referral center.

Patients
Fifty-six consecutive immunocompromised children (ie, the prospective group) with laboratory evidence consistent with acute or recently acquired CMV infection, which was defined as CMV cultured from the blood, urine, nasopharynx, or biopsy specimen, recent seroconversion, a 4-fold increase in the CMV antibody titer, or an antibody titer of 1:512 or greater. Ninetythree immunocompromised children (ie, the retrospective group) with acute CMV or previous CMV exposure, which was defined as a CMV titer of 1:4 or greater and less than 1:512.

Main Outcome Measure
Occurrence of CMV retinitis.

Results
Cytomegalovirus retinitis developed in 3 children in the prospective group and in 4 children in the retrospective group. The causes of immunosuppression were severe combined immunodeficiency syndrome (n=2), severe combined immunodeficiency syndrome status post bone marrow transplantation (n=1), acquired immunodeficiency syndrome (n=l), and acquired immunodeficiency syndrome status post bone marrow transplantation for leukemia (n=1), renal transplantation (n=l), and chemotherapy for leukemia (n=l). Cytomegalovirus retinitis was associated with a positive CMV culture result from the urine (P=.03) or nasopharynx (P<.001) in the retrospective group. In the retrospective group, one child with congenital CMV infection and CMV retinitis was excluded from analysis because laboratory tests for CMV were not obtained prior to ganciclovir therapy.

Conclusions
Cytomegalovirus retinitis is uncommon in children compared with adults; it occurred in 5% of the children in our series. A screening ophthalmologic examination should be considered in immunocompromised children with positive CMV laboratory results, particularly positive results of urine or nasopharynx cultures.

Arch Pediatr Adolesc Med. 1996;150:1186-1192