Seasonal trends in month of diagnosis have been reported for childhood acut
e lymphoblastic leukaemia (ALL) and non-Hodgkin's lymphoma (NHL). This seas
onal variation has been suggested to represent an underlying viral aetiolog
y for these malignancies. Some studies have shown the highest frequency of
diagnoses in the summer months, although this has been inconsistent. Data f
rom the Children's Cancer Group and the Pediatric Oncology Group were analy
sed for seasonal incidence patterns. A total of 20 949 incident cancer case
s diagnosed in the USA from 1 January 1989 through 31 December 1991 were av
ailable for analyses. Diagnosis-specific malignancies available for evaluat
ion included ALL, acute myeloid leukaemia (AML), Hodgkin's disease, NHL, rh
abdomyosarcoma, neuroblastoma, retinoblastoma, osteosarcoma, Wilms' tumour,
retinoblastoma, Ewings' sarcoma, central nervous system (CNS) tumours and
hepatoblastoma. Overall, there was no statistically significant seasonal va
riation in the month of diagnosis for all childhood cancers combined. For d
iagnosis-specific malignancies, there was a statistically significant seaso
nal variation for ALL (P = 0.01; peak in summer), rhabdomyosarcoma (P = 0.0
3; spring/summer) and hepatoblastoma (P = 0.01; summer); there was no seaso
nal variation in the diagnosis of NHL. When cases were restricted to latitu
des greater than 40 degrees ('north'), seasonal patterns were apparent only
for ALL and hepatoblastoma. Notably, 33% of hepatoblastoma cases were diag
nosed in the summer months. In contrast, for latitudes less than 40 degrees
('south'), only CNS tumours demonstrated a seasonal pattern (P = 0.002; wi
nter). Although these data provide modest support for a summer peak in the
diagnosis of childhood ALL, any underlying biological mechanisms that accou
nt for these seasonal patterns are likely complex and in need of more defin
itive studies. (C) 1999 Cancer Research Campaign.