Results
We initially identified 2080 studies and 80 studies met the selection criteria. For flowchart of the study screening and inclusion, see Figure 1. The authors of these 80 studies had all published several studies originating from the same childhood cancer cohort. Including only the first study published on the cohort resulted in 16 studies each representing different childhood cancer cohorts (Table 1): one Nordic study (Denmark, Norway, Sweden, Iceland and Finland) [11]; three in Norway (two based in a single center [12, 13] and one national [14]); five in Sweden (all in a single center [15-19]); six in Finland (three in a single center [20-22] and three national cohorts [23-25]); and one in a single center in Iceland [26]. The full study sample based on the respective 16 cohorts identified included 879 childhood cancer survivors comprising 341 females and 538 males.
Recruitment procedures of the person who contacted the participant (physician, nurse or other) and how the participant was approached (postal letter, mail or telephone) were not consistently reported. In six cohort studies, a letter was send to the participants, whereas in six other cohort studies a physician was the direct contact person. Information on the female-male ratio among the invited participants, study drop-outs, and study non-participants was not described in the studies. Only sex distribution of the accepted participants was available. Two male-only studies investigated testicular function, whereas 14 studies comprised both males and females.
Clinical examinations included training programs, endocrinologic or audiologic evaluations, radiologic imaging and neurocognitive evaluations. The number of participating females was highest in patients with solid tumors (19 females of 34 participants, 55.8%) and mixed diagnosis (42 females of 77 participants, 54.5%) compared to those with hematologic malignancies (174 females of 358 participants, 48.6%) and central nervous system tumors (106 females of 223 participants, 47.5%). Further characteristics of the participants in the included studies are displayed in Table 1.
SPR ranged from 27 to 100% (median 72.5%) (Figure 2). The lowest SPR was registered by Järvela et al. (27%) [27], who investigated physical activity with a home-based training program for survivors of acute lymphoblastic leukemia. The highest SPR being 100% was reported by Einarsson et al. [19] and Einar-Jon et al. [26] both performing audiometric evaluation of survivors receiving platinum-based chemotherapy.
The age at diagnosis ranged from 3.0 to 14 years (median 6.0). The median age at diagnosis was lowest for studies investigating late effects of acute myeloid leukemia (median 3.0 years, range 0-15) [11]. The highest median age at diagnosis was 14 years (range 0-18) and at study entry 32 years (range 19-55) observed in an interview study involving clinical examination of late effects after malignant lymphoma [28]. The lowest age at study entry was observed in a study of neurocognitive late effects after acute lymphoblastic leukemia (median 11.5 years, range 8.4-15.3) [12].