Background/purpose: For children with probable mediastinal lymphoma and a high risk of cardiorespiratory morbidity, many centers recommend delaying the diagnostic biopsy for 24 to 48 hours while corticosteroids are administered to reduce tumor size and morbidity. This study was undertaken to determine the effect of preoperative steroid use on the accuracy of the pathological diagnosis and incidence of perioperative cardiorespiratory morbidity.
Methods: From 1988 to 1998, 86 children were treated for mediastinal lymphoma. Twenty-three received steroid before biopsy (study group) because of clinical evidence of respiratory compromise, and the remaining 63 served as controls. Clinical parameters, steroid use, and detailed pathological findings obtained at initial and subsequent biopsies were reviewed. Steroid treatment was considered to have had an adverse effect on the pathological diagnosis if (1) a definitive diagnosis was delayed more that 1 month, (2) a definitive diagnosis could not be made, or (3) the extent of disease could not be staged with certainty.
Results: Steroid treatment had an adverse effect on the pathological diagnosis in 5 of 23 (22%) children: 1 diagnostic delay, 3 failures of a definitive diagnosis, and 1 possible failure of staging. A definitive diagnosis was made in all control patients. Perioperative survival was 100% in both groups. At biopsy, only 3 children in the steroid treatment group and 2 children in the control group had moderate, nonfatal cardiorespiratory instability. Parameters of steroid use among children who had inaccurate pathological diagnoses or cardiorespiratory morbidity were not significantly different from those who did not.
Conclusions: Steroid treatment before biopsy of mediastinal lymphoma may adversely affect the pathological accuracy or cause a delay in definitive diagnosis in a minority of cases. The dose and duration of steroid use was not related to outcome. Prebiopsy steroid appears to minimize the likelihood of cardiorespiratory morbidity in high-risk patients.