A controlled comparison of ovarian response to controlled stimulation in first generation Asian women compared with white Caucasians undergoing in vitro fertilisation


Correspondence: Dr H. Lashen, Department of Obstetrics and Gynaecology, Solihull Hospital, Lode Lane, Solihull, West Midlands B91 2JL, UK.


Objective To compare ovarian response to controlled stimulation among Asian women from the Indian sub-continent and white Caucasian women undergoing in vitro fertilisation (IVF).

Design Nested case-control study.

Setting Assisted Conception Unit, Birmingham Women's Hospital.

Sample One hundred and eight first generation Asian patients (born in the Indian sub-continent) and 216 white Caucasian controls, all of whom received IVF treatment in the period 1994 to 1997, were selected for the study. The two groups were matched for age to within one year, early follicular phase follicle stimulating hormone, indication for treatment, gonadotrophin dose and year of treatment. The outcome of treatment was not known when the controls were selected.

Results There was no statistically significant difference between the two groups in the duration of stimulation, egg number, number of embryos produced, fertilisation rate, clinical pregnancy rate, miscarriage rate, cycle cancellation rate and implantation rate.

Conclusion Under the same IVF protocol Asian women's response to controlled ovarian stimulation and IVF outcome are comparable to their white Caucasian peers.


The epidemiology of diseases and the response to treatment can vary in different parts of the world. Whether this difference is genetically or environmentally determined is not entirely clear. The relatively large number of immigrants to North America and Europe has presented the health professions with a challenge as well as a research opportunity. For example, by migrating to the USA, Japanese and Chinese women increase their risk of breast cancer, but their risk remains less than that of women born in the USA, including second generation Japanese and Chinese women1.

In the United Kingdom, the reproductive potential of Asian women (from the Indian sub-continent) has been studied with regard to perinatal mortality and fetal birth-weight2,3. However, the fertility of this ethnic group has received little attention.

Anand-Kumar et al.4 drew the attention to a possible effect of ethnic origin on in vitro fertilisation (IVF) out-come when they reported that Indian women have a poorer performance in IVF compared with white women. They suggested a relative lack of response to ovarian stimulation as the possible reason for such a difference. On the other hand, Ratnam and Devendra5 reported IVF results comparable to those achieved in the western world. Neither of these reports came from a comparative study. In a controlled comparative study carried out in Oxford, it was found that under the same IVF regime, second generation Asian women in the UK performed worse than white women at the ovarian stimulation stage6.

Asian women represent approximately 10% of patients in our unit, and the majority of them are first generation immigrants to the UK, mainly from Pakistan and the Punjab. Therefore we had an excellent opportunity to review their response to ovarian stimulation and IVF outcome compared with white Caucasian women, so that they could be given information relevant to their chances of success prior to starting IVF treatment.


The IVF data, including all the cycles in which gonadotrophin stimulation is started, were prospectively entered in a register which was later entered into a database computer programme. We reviewed the IVF database over the period from 1994 to 1997. The name, ethnic origin, age, indication for treatment, day 3 follicle stimulating hormone, and daily gonadotrophin dose of all the patients treated during the study period were retrieved. We included only each woman's first cycle of treatment in the study. Ninety-eight percent of the women from the Indian subcontinent treated during this period were first generation immigrants to the UK, and they were selected for the study. For each Asian patient two matched white Caucasian controls were obtained by selecting the next successive patient in the database until a match was found. The matching criteria were: age to within one year, day 3 follicle stimulating hormone ± 1 IU/L, daily gonadotrophin dose, and the indication for IVF (male factor and non-male factor including failure of donor insemination).

The total gonadotrophin dose, duration of ovarian stimulation, number of eggs, fertilisation rate, number of embryos, cycle cancellation rate, clinical pregnancy rate and implantation rate were the main outcome measures compared between the study and control groups. The incidence of primary and secondary infertility, duration of infertility and the incidence of polycystic ovaries were compared between the study and the control groups. These data were not retrieved until the selection of the control group had been carried out, to avoid any bias in selecting the control group. Patients who required micro-manipulation were not included in the study.

IVF protocol

The protocol was consistent during the study period. A long protocol of pituitary down-regulation was used in all the patients. Intra-nasal nafarelin acetate (Synarel, Searle, UK) 200 μg eight hourly was started on cycle day 21, for at least two weeks. The patient was then seen for a pelvic ultrasound scan to assess the response to down-regulation. An endometrial thickness of < 4 mm and inactive ovaries indicated successful down-regulation, and ovarian stimulation with gonadotrophin was then started. The daily dose was decided according to the patient's age: women aged < 35 years received 3 ampoules (225 IU), those between 35 and 40 received 4 ampoules (300 IU) and those > 40 received 6 ampoules (450 IU). Ovarian stimulation with gonadotrophin was continued in addition to nafarelin acetate until human chorionic gonadotrophin was administered. Follicular development was monitored using vaginal ultrasound scanning. When the leading follicle measured 20–22 mm in diameter (with two more follicles ≥ 16 mm in diameter), a single dose of 10,000 units of human chorionic gonadotrophin was administered intramuscularly and oocyte retrieval was carried out transvaginally under ultrasound guidance 36 h later. Embryo transfer was carried out 48 h or 72 h later.

If a poor response to stimulation was obtained the cycle was cancelled. Poor response was defined as < 3 mature follicles measuring ≥ 16 mm in diameter. However, cycle cancellation for fear of ovarian hyperstimulation syndrome was not our practice. If ovarian hyperstimulation syndrome was predicted, all the embryos produced were frozen and embryo transfer was abandoned.

A urinary pregnancy test was performed on the first morning sample two weeks after embryo transfer. If the pregnancy test was positive a transvaginal ultrasound scan was carried out two weeks later to document the presence of fetal cardiac activity to diagnose a clinical pregnancy.

Statistical analysis

This was carried out using Minitab for Windows statistical package (Minitab Inc, Pennsylvania, USA). The results are presented as mean (SEM) and median. Due to the non-Gaussian distribution of data, the non-parametric Mann-Whitney U test was used to compare the quantitative variables. The χ2 test and logistic regression analysis were used to compare the categorical data as appropriate. P < 0.05 was considered significant.


During the study period 1040 women underwent IVF treatment, from whom the study group and their controls were selected (108 first generation Asian patients and 216 white Caucasian). 58/108 (53.7%) of the Asian patients were Pakistani and 34/108 (31.5%) were from the Punjab, giving 85% from the northern part of the Indian subcontinent, representing a homogeneous study group. The matching criteria and the outcome measures between the two groups are summarised in Tables 1 and 2, respectively. There was no significant difference between the Asian women and the white controls in any of the outcome measures (Table 2). However, a significantly longer duration of infertility (P < 0.001), and a higher incidence of polycystic ovaries (odds ratio = 2.7, 95% CI 1.25–5.8, P= 0.01) were noticed in the Asian group compared with the control group. The median duration of infertility (95% CI) in the Asian and control groups was 7.5 (6.5:8.5) and 5 (4.55) years, respectively.

Table 1.  Matching criteria of Asian women and white Caucasian controls. FSH = follicle stimulating hormone; NS = not significant.
 Asian (n= 108)Caucasian (n= 216) 
 Mean (SEM)MedianMean (SEM)MedianP
Age (years)32.3 (0.5)3232.3 (0.3)32.5NS
Day 3 FSH (IU/L)6.9 (0.2)6.77.4 (0.2)7.2NS
Daily dose (IU/L)275.2 (10.5)262.5275.2 (10.5)262.5NS
Table 2.  Ovarian response and in vitro fertilisation outcome. PR = poor response; NS = not significant (P > 0.05).
 Asian (n= 108)Caucasian (n= 216) 
 Mean (SEM) or n [%]MedianMeans (SEM) or n [%]MedianP
Total dose (IU/L)3253 (138)29252980 (69.3)2812NS
Days stimulation11.8 (0.1)1211.6 (0.1)11.5NS
Cycle cancellation (for PR)9 [8]18 [8]NS
Eggs11.6 (0.8)1111.9 (0.5)11.5NS
Fertilisation rate0.5 (0.03)0.50.6 (0.02)0.6NS
Embryos5 (0.5)4.55.7 (0.3)5NS
No. of embryos transferred2.4 (0.08)2.52.4 (0.05)2.5NS
Clinical pregnancy/cycle17 [16]49 [22.6]2.5NS
Implantation rate13%17%NS


Patients' performance in IVF can be influenced by many variables7. The effect of ethnic origin on IVF out-come has received little attention, despite some controversial reports. In this study, we found that the ovarian response and IVF outcome of the first generation Asian women were comparable to white Caucasian women, which contradicts what has been previously reported4,6. However, the first study4 was not comparative, and the sample size in the second study by Mahmud et al.6 was much smaller than in our study. Besides, Mahmud et al. failed to include the basal follicle stimulating hormone level of 17/44 (39%) of the Indian patients. The possibility of high levels in these patients cannot be ruled out, and this can have a significant impact on the outcome in such a small study group. In agreement with Mahmud et al., we found that the difference between clinical pregnancy and implantation rates failed to reach statistical significance. However, the longer duration of infertility in the Asian group may explain the lower pregnancy rate in this group. A longer duration of infertility in the Asian group was previously reported by Mahmud et al.6. The reason for this confirmed trend is not clear and needs to be investigated. Perhaps a better understanding of Asian community culture may explain this finding.

In agreement with the general impression of most clinicians, we found a higher incidence in polycystic ovaries among the Asian women compared with the controls. However, a larger epidemiological study, preferably in the general population, both fertile and infertile, is needed to verify this finding.

We believe that our study is the largest controlled study comparing the outcome of IVF between Asian and white Caucasian women, and the first of its kind to study the IVF outcome in first generation Asian women in the UK. The lack of previous similar studies precluded our estimating the sample size required to achieve a 90% statistical test power at a confidence level of 0.05. Estimating the power in retrospect is very complicated in this study, due to the unequal size of the study and control groups. However, using the 6% difference in the pregnancy rate in this study to estimate a sample size for a prospective study indicates that 130 patients in each arm should achieve a 90% power at a 0.05 confidence level.