Skip to main content
. 2020 Nov 3;23(4):341–356. doi: 10.1111/bdi.13016

Table 1.

Details of the studies which report a trimodal age‐at‐onset distribution in bipolar disorder

Study N Country Diagnosis Recruitment Definition of age at onset Method of determining AAO Mean age of sample at study entry (SD) Early onset Mid‐onset Late onset
Upper age limit Mean (SD), % Lower and upper age limits Mean (SD), %

Lower

age

limit

Mean (SD), %
Azorin et al (2013) 1082 France DSM‐IV BPI The EPIMAN II Mille study, a multi‐centre naturalistic study conducted in 19 French medical centres Age at which the patient first met the Research Diagnostic Criteria for an affective episode

Medical records.

Structured interviews with patients and relatives.

42.9 (13.7) 20 18.6 (2.1), 19% 21‐29 24.3 (5.3), 38.9% 30 36.7 (10.8), 42%

Bellivier et al (2001)

211 France DSM‐IV BPI Consecutive inpatients and outpatients in France Age at which DSM‐IV criteria for an affective episode was first met

Medical records.

Diagnostic Interview for Genetic Studies

42.4 (14.8) 16.9 (2.7), 41.4% 26.9 (5.0), 41.8% 46.2 (8.0), 16.6%
Bellivier et al (2003) 579

France,

Switzerland

Germany

Ireland

DSM‐IV BPI

Inpatients and outpatients across four countries

Age at which DSM‐IV criteria for an affective episode was first met

Medical records.

Diagnostic Interview for Genetic Studies

Not reported 17.4 (2.3), 27.9% 25.1 (6.2), 50.1% 40.4 (11.3), 21.9%
Bellivier et al (2014) 5891

Europe and USA

DSM‐IV BPI

Recruited for genetic, pharmacological and observational studies across 18 sites in Europe (N = 3616, incl. participants from the EMBLEM study) and from the Stanley Centre Bipolar Registry in the USA (N = 2275)

Age at which DSM‐IV criteria for an affective episode were first met Semi‐structured interview 44.0 (13.2) Europe
19 (2.7), 24.8% 27.2 (6.3), 50.7% 41.8 (10.7), 24.5%
40.8 (11.7) USA
14.5 (4.9), 63.0% 26.5 (7.6), 28.5% 39.5 (12.5), 8.5%
Biffin et al (2009) 162 Australia DSM‐IV BPI Recruited as part of the Bipolar Comprehensive Outcome Study (BCOS) in Melbourne, Australia Self‐reported age at which episode of mania or depression first met diagnostic criteria Questionnaire developed by the research team

Early:38.7 (12.6)

Mid: 43.7 (12.6)

Late: 58.9 (11.5)

15.5 (2.7), 44.4% 26.1 (4.8) 48.1% 50.6 (9.0), 7.4%
González‐Pinto et al (2009) 169 Spain DSM‐IV BPI Inpatients and outpatients who were receiving treatment in Alava, a Spanish province. The age at first treatment for an affective disorder Medical records. Semi‐structured SCID‐P interview. Emergency service records. Interviews with relatives. 46.0 (16.0) 18.2 (2.0), 34.0% 26.1 (5.5), 44.0% 50.9 (9.1), 22.0%
Hamshere et al (2009) 1369 UK DSM‐IV BPI Large‐scale genetic epidemiological study. Recruited via community mental health teams, general practitioner surgeries and patient support organisations across the UK. Age at first impairment due to an affective episode according to self‐report Medical records. Schedules for Clinical Assessment in Neuropsychiatry (SCAN). Age range: 6 to 73 years 22 18.7 (3.7), 47.1% 25‐37 28.3 (5.5), 38.8% 40 43.3 (9.1), 14.3%
Lin et al (2006) 211 USA

DSM‐III‐R

BP‐I

NIMH Genetics Initiative for Bipolar Disorder 85 Self‐reported age at which episode of (hypo)mania or depression first met diagnostic criteria Medical records. Diagnostic Interview for Genetic Studies. Age range: 0 to >61 years 21 16.6 (5.1), 79.7% 22‐28 26.0 (1.4), 7.2% 28 34.7 (6.6), 13.1%

Manchia et al (2008)

181 Sardinia

RDC‐BPI

Recruited from the Lithium Clinic of the Clinical Psychopharmacology Centre, University of Cagliari, Italy

Age at first reliably diagnosed (hypo)manic or depressive episode

Medical records. Semi‐structured interview 42.8 (14.8) 20

18.1 (2.3),

36.0%

21‐33

24.3 (5.3),

39.0%

34

41.0 (11.5),

25.0%

Nowrouzi et al (2016) 194 Canada

DSM‐III‐R or DSM‐IV

BPI

BPII

Recruited from four clinical sites across Ontario, Alberta and British Columbia, Canada Unknown Unknown

25.2 (9.51),

range 14‐65 years

18.0 (2.9), 69.0% 28.7 (3.5), 22.0% 47.3 (7.8), 9.0%
Ortiz et al (2011) 379 Canada

DSM‐IV or RDC

BPI

BPII

Recruited through the Maritime Bipolar Registry, a community‐based

project in the Maritime Provinces of Canada 86

Age at which DSM‐IV criteria for an affective episode was first met (according to medial case notes and interviews) Medical records. Schedule for Affective Disorders and Schizophrenia, Lifetime version 50.1 (12.7) 19

15.5 (2.0),

29.0%

20‐31 22.8 (4.6), 37.1% 32 36.1 (10.1), 33.4%
Severino et al (2009) 300 Italy

RDC BPI

BPII

Schizoaffective bipolar manic type

Outpatients at the Lithium Clinic of the Clinical Psychopharmacology Centre, University of Cagliari, Italy Age at first reliably diagnosed (hypo)mania or depression according to RDC criteria (using medical records) Medical records. Semi‐structured interview 42.9 (14.8) 22 18.5 (2.6), 43.0% 23‐37 27.5 (6.1), 42.0% 38 43.0 (10.8), 15.0%
Tozzi et al (2011) 964 UK and Canada

DSM‐IV or ICD‐10

BP‐I

BP‐II

Recruited across three sites: Toronto (Canada) at the Centre for Addiction and Mental Health, London (UK), at the Institute of Psychiatry, and Dundee (UK) at the University of Dundee Self‐reported age at which episode of mania or depression first met diagnostic criteria Schedules for Clinical Assessment in Neuropsychiatry (SCAN) interview

47.2 (12.1),

range 18‐84 years

24 16.1 (4.2), 64.0% 25 25.4 (2.5), 6.0% 26 32.2 (9.5), 30.0%
Grigoroiu‐Serbanescu et al (2014)a 1857

Germany

Poland

Romania

DSM‐IV BPI Consecutive inpatients recruited at three sites: Age at which DSM‐IV criteria for an affective episode were first met Medical records. Semi‐structured interview with patients and relatives 43.4 (13.4) Romania
17.6 (3.2), 43.0% N/A 20‐21 29.9 (8.2), 57.0%
17.3 (2.8) 33.0% 25.6 (6.3), 46.0% 40.9 (5.3), 21.0%
44.0 (13.4) Germany
20.7 (6.0), 67.0% N/A 25 38.4 (6.5), 33.0%
19.3 (5.5), 46.0% 28.5 (7.1), 41.0% 45.4 (4.7) 13.0%
45.0 (14.1) Poland
20.47 (3.91), 65% N/A 24‐25 33.57 (9.12), 35%
20.7 (3.7), 44.0% 33.0 (6.1), 45.0% 49.0 (5.3), 11.0%

Age bounds for the subgroups are provided. Numbers reported to one decimal place.

a

Two component and three component models fitted the data equally well.