Background: Lifetime rates of depression are distinctly higher in women reflecting both real and artefactual influences. Most prevalence studies quantifying a female preponderance have examined severity-based diagnostic groups such as major depression or dysthymia. We examined gender differences across three depressive sub-type conditions using four differing measures to determine whether any gender differences emerge more from severity or symptom prevalence, reflect nuances of the particular measure, or whether depressive sub-type is influential.

Methods: A large clinical sample was recruited. Patients completed two severity-weighted depression measures: the Depression in the Medically Ill 10 (DMI-10) and Quick Inventory of Depressive Symptoms-Self-Report (QIDS-SR) and two measures weighting symptoms and illness correlates of melancholic and non-melancholic depressive disorders - the Severity of Depressive Symptoms (SDS) and Sydney Melancholia Prototype Index (SMPI). Analyses were undertaken of three diagnostic groups comprising those with unipolar melancholic, unipolar non-melancholic and bipolar depressive conditions.

Results: Women in the two unipolar groups scored only marginally (and non-significantly) higher than men on the depression severity measures. Women in the bipolar depression group, did however, score significantly higher than men on depression severity. On measures weighted to assessing melancholic and non-melancholic symptoms, there were relatively few gender differences identified in the melancholic and non-melancholic sub-sets, while more gender differences were quantified in the bipolar sub-set. The symptoms most commonly and consistently differentiating by gender were those assessing appetite/weight change and psychomotor disturbance.

Conclusion: Our analyses of several measures and the minimal differentiation of depressive symptoms and symptom severity argues against any female preponderance in unipolar depression being contributed to distinctly by these depression rating measures. Our analyses indicated that gender had minimal if any impact on depression severity estimates. Gender differences in depressive symptoms and severity were more distinctive in bipolar patients, a finding seemingly not previously identified or reported.

Limitations: The study had considerable power reflecting large sample sizes and thus risks assigning significant differences where none truly exist, although we repeated analyses after controlling for the type I error rate.