Letters

For appreciable ethnic differences in false positive rates, use different cut off points

BMJ 1996; 312 doi: https://doi.org/10.1136/bmj.312.7037.1041 (Published 20 April 1996) Cite this as: BMJ 1996;312:1041
  1. P J Holloway,
  2. S Bulusu
  1. Senior biomedical scientist Consultant chemical pathologist Chemical Pathology Department, Newham General Hospital, London E13 8RU

    EDITOR,—Lucy Gilbert and colleagues report ethnic differences in biochemical screening for Down's syndrome and the relative ineffectiveness of the programme in Indian Asians.1 At national meetings of the Association of Clinical Biochemists we presented our findings on the ethnic differences in maternal serum (alpha) fetoprotein and free β human chorionic gonadotrophin concentrations2 and the differences in false positive rates in the various groups, which remain unchanged even after the use of “ethnospecific” medians.3

    The false positive rate was highest in Afro-Caribbean women (9.2%; n=185) and lowest in white women (4.6%; n=328), with that in South Asian women being between the two (6.3%; n=315). Since then we have re-examined our database as we have acquired considerably more data. Contrary to our previous findings, while the false positive rate remains highest in Afro-Caribbean women, it is slightly lower in South Asian women than in white women. The false positive rates remain essentially unchanged, as before, after use of ethnospecific medians (table 1). The reason for this difference in the false positive rate in the South Asian women is not clear and could be the considerably smaller numbers we had at the time.

    Table 1

    Ethnic differences in biochemical screening for Down's syndrome*

    View this table:

    Reports suggesting that the incidence of Down's syndrome is low in South Asians, even after migration,4 seem consistent with our findings. We remain unconvinced that the detection rate (0/2) found in “Indian Asians” by Gilbert and colleagues is sufficient for it to be concluded that screening is ineffective in these women. Our data do not support this view. The low incidence of trisomy in South Asian women, if confirmed, might be a more convincing argument for abandoning biochemical screening in these women. More information about the true incidence of trisomy in South Asians in Britain is necessary before any major changes in screening strategies are undertaken.

    The differences in false positive rates are probably due to different curve parameters for (alpha) fetoprotein and human chorionic gonadotrophin in normal and Down's syndrome pregnancies in different ethnic groups. We recommend that laboratories serving multiracial populations should examine the false positive rates separately in the different ethnic groups as the average rate might be misleadingly low; when there are appreciable ethnic differences in the false positive rates they should consider using different cut off points, particularly in Afro-Caribbean women, to avoid subjecting these women to unnecessary anxiety and invasive investigations.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    View Abstract