Background The effectiveness of in-vitro fertilisation (IVF) treatment
depends both on the overall success rate in the treating clinic and o
n the characteristics of the couple seeking treatment. Since 1991, the
Human Fertilisation and Embryology Authority (HFEA) has been collecti
ng information on all IVF cycles carried out in the UK. This database
has been analysed to identify the factors that affect the outcome of t
reatment. Methods All IVF treatment cycles and outcomes registered bet
ween August, 1991, and April, 1994, were identified (52 507). Cycles t
hat involved gamete or embryo donation, frozen embryo transfer, or mic
romanipulation and unstimulated cycles were excluded. Thus, 36 961 cyc
les (70% of those registered) were included in the analysis. The main
outcome measure was livebirth rate per cycle started, The relation bet
ween age and outcome was investigated by fitting of different fraction
al polynomials of age with logistic regression models. All other facto
rs were analysed by logistic regression with age included in the model
. Findings The overall livebirth rate per cycle of treatment was 13.9%
. The highest livebirth rates were in the age-group 25-30 years; young
er women had lower rates and there was a sharp decline in older women.
At all ages over 30, use of donor eggs was associated with a signific
antly higher livebirth rate than use of the woman's own eggs, hut ther
e was also a downward trend in success rate with age (p=0.04. After ad
justment for age, there was a significant decrease in livebirth rate w
ith increasing duration of infertility from 1 to 12 years (p<0.001). T
he medical indication for treatment had no significant effect on the o
utcome. Previous pregnancy and livebirth significantly increased treat
ment success. The possibility of success decreased with each IVF treat
ment cycle. Interpretation We were able to identify by logistic regres
sion the factors that significantly affect the outcome of IVF treatmen
t, and to measure the magnitude of that effect. These factors should b
e taken into account in assessment of IVF results. After allowance for
background clinic success rates, these factors can be used to predict
outcome in individual cases.