The prognostic impact of karyotypic subgroups in myelodysplastic syndromes is strongly modified by sex

The prognostic impact of karyotypic patterns in a consecutive series of 389 adult myelodysplastic syndromes (MDS) was investigated. Time period did not significantly influence the survival times. In the analyses, the MDS cases were subdivided into the cytogenetic subgroups used in the International Prognostic Scoring System, i.e. favourable [-Y, del(5q) or del(20q) as single aberrations or normal karyotype, n = 241], poor [-7, del(7q), der(1;7) or complex karyotypes, i.e. greater than or equal to three abnormalities, n = 89] and intermediate (other aberrations, n = 59). The survival times correlated well with the prognostic subgroups, confirming that the cytogenetic. classification was valid. Expressed as hazard ratios (HRs), with the favourable subgroup as the reference, the intermediate and poor subgroup HRs increased to 1.5 (95% confidence interval, 1.1-2.1) and 3.2 (2.4-4.1) respectively. Sex, age, morphological subtype and smoking habits significantly modified this prognostic impact. Shorter survival was detected for men in the favourable and the intermediate subgroups, but not in the poor prognosis subgroup. Using women in the favourable subgroup as the reference and adjusting for age, the HR for men was 1.6 (1.2-2.1) in the favourable subgroup. Adjusting for smoking habits as well decreased the HR to 1.4 (1.1-2.0) and, when also excluding cases with del(5q) as the sole anomaly, no significant difference could be discerned [HR 1.2 (0.9-1.6], suggesting that the better outcome for women in the favourable subgroup was mainly as a result of the '5q-syndrome' and to smoking habits. In the intermediate subgroup, the corresponding HRs were 3.0 (1.5-6.0) when adjusted for age and 2.7 (1.3-5.5) when also adjusted for smoking habits. Different survival times between men and women have never previously been reported for this MDS group. Although it remains to be elucidated whether environmental and/or constitutional factors cause the observed sex-related difference, these observations have obvious clinical ramifications, not least in designing and evaluating therapy protocols.