Papers

Ethnic differences in the outcome of serum screening for Down's syndrome

^a Department of Obstetrics and Gynaecology, Oldham NHS Trust, Oldham OL1 2PN, ^b Medical Care Research Unit, University of Sheffield, Sheffield S10 2RX, ^c Department of Clinical Pathology, Oldham NHS Trust, ^d Regional Cytogenetics Department, St Mary's Hospital, Manchester M13 0JH

Correspondence to: Mrs L Gilbert, Women's Hospital, Doncaster Royal Infirmary, Doncaster DN2 5LT.

Serum screening for Down's syndrome was introduced in Oldham health district in January 1991 to reduce the false positive rate associated with screening based on age alone and to improve the detection rate.¹ A retrospective analysis was undertaken to evaluate whether these objectives were achieved throughout the entire obstetric population in Oldham.

Subjects, methods, and results

All women booked at the Royal Oldham Hospital and the community clinic before 18 weeks' gestation are counselled and offered serum screening by the booking midwife, who also records their ethnic origin. The gestation is confirmed by an ultrasound scan. (alpha) Fetoprotein and intact human chorionic gondatrophin values are measured by enzyme immunoassay. The risk for Down's syndrome is calculated from the two analyte results (adjusted for weight) and the maternal age.² Women who have a positive result (estimated risk of 1 in 300 or higher of a fetus with Down's syndrome) are counselled by a senior obstetrician and offered an amniocentesis. A trained interpreter is used if necessary.

Between 1 February 1991 and 31 October 1993, of the 9217 women offered serum screening 1655 (18%) were classified as "Indian Asians"--that is, from the Indian subcontinent. Pakistani and Bangladeshi women formed the largest subgroups, 956 (10.3%) and 605 (6.5%) respectively, with Indians and East African Indians accounting for the rest. Other ethnic minorities were counted together with the white population because of their small numbers: West Indian 51 (0.5%), Chinese 22 (0.2%), African 6 (<0.1%), Arab 9 (<0.1%), Mediterranean 6 (<0.1%). The outcomes of the pregnancies in the study population were determined from the obstetric, paediatric, and regional cytogenetics databases; the last includes all antenatal and postnatal karyotyping carried out in the North West region. In addition, all children with Down's syndrome known to social workers were identified. Proportions within ethnic groups were compared by using ² tests and 95% confidence intervals³ calculated for proportions and for the relative risks of positive screening.

Among the white and other women, serum screening improved the detection rate for Down's syndrome from a maximum of 25% (3/12) for a 7.8% amniocentesis rate if selection for amniocentesis had been based on age alone to 42% (5/12) for a 3.5% amniocentesis rate (table). It also extended the opportunity for screening to women under 35, in whom 80% of Down's pregnancies in these women occurred. In women of Asian origin, however, there was a significantly higher false positive screening rate (12.3% v 3.4%), seen both in those aged 35 and over and in those aged under 35, with no evidence of a higher incidence of Down's syndrome or an improved detection rate. The failure to detect either of the two babies with Down's syndrome born to serum screened Asian women was not due to their significantly lower uptake of amniocentesis. Both babies were born to women who gave a negative result on screening.

Comment

The apparent failure of serum screening in the Asian population may be due to ethnic influences (cultural and religious beliefs, uncertainty about age) as well as racial differences. The age specific risk estimates and the medians of the biochemical markers used in the United Kingdom are derived from largely white populations and extrapolated to other populations on the assumption that "there is no reason to believe that they cannot be generally applied."¹ Yet racial differences have been found in the age specific rates for Down's syndrome⁴ and in the medians of the biochemical markers.⁵

Outcome of antenatal Down's syndrome screening. Values are numbers (percentages; 95% confidence intervals) unless
otherwise indicated
-----------------------------------------------------------------------------------------------------------------------------------------------
                                       All women           White and other             Indian Asians
-----------------------------------------------------------------------------------------------------------------------------------------------
Total eligible for screening          9217                 7562                        1655
Uptake of serum screening             7953 (86.3)          6765 (89.5)                 1188 (71.8)
Women screened positive                382 (4.8)            236 (3.5)                   146 (12.3)
 Relative risk                                                                          3.5 (2.9 to 4.3)
Amniocentesis uptake                   179 (48)             142 (60)                     37 (27)
 12                                                                   43.9 (P<0.001)
Detection rate for Down's syndrome    5/14 (36)            5/12                         0/2
                                                           (42; 13 to 72)               (0; 0 to 84)
Positive predictive value             1:76 (5/382)              1:47                      0:146
No of cases of Down's syndrome:
 Identified antenatally                    6*                   6*                          0
 Identified postnatally                   12+                   9+                          3+
Rate per 1000 births                       1.8++                1.8++                       1.8
-----------------------------------------------------------------------------------------------------------------------------------------------
*Includes one case of Down's syndrome identified by primary amniocentesis; details available from the authors on request.
+Includes the Down's syndrome births missed by serum screening and those occurring in unscreened women.
++After a 23% reduction in antenatally diagnosed and terminated cases to allow for the increased fetal death rate in pregnancies with Down's
syndrome diagnosed by amniocentesis but not ending in a termination.1

This study suffers from the disadvantage that it is retrospective and confined to one health district. Nevertheless, it highlights the need for outcome data on serum screening to be collected by ethnic group, so that problems experienced by specific groups can be recognised and addressed. To date none of the published studies on serum screening has looked at the outcome by ethnic group. Ethnic minorities form a significant proportion of the obstetric population of some health districts, and to justify serum screening in these groups it must be based on validated risk estimations and be shown to be effective.

We thank I Basnett (Camden and Islington Health Authority); D E Mutton (Wolfson Institute of Preventive Medicine, St Bartholomew's Hospital); C Tyler (Public Health Resource Centre, Oldham, Rochdale and West Pennine Health Authorities); and, from the Oldham NHS Trust, M Scholes and S Cavanagh (midwifery department), J McAvedy (special care baby unit), R Subramanium and B Plant (ultrasound department), and T Windle and S Monteith (general manager and patient services manager).

Funding: No additional funding.

Conflict of interest: None.

1. Cuckle HS, Wald NJ. Screening for Down's syndrome. In: R J Lilford, ed. Prenatal diagnosis. London: Butterworth, 1990:67-92.
2. Reynolds TN, Penney MD. The mathematical basis of multivariate risk screening with special reference to screening for Down's syndrome associated pregnancy. Ann Clin Biochem 1989;26:26-37.
3. Gardner MJ, Altman DG, eds. Statistics with confidence. London: BMJ, 1989.
4. Rodgers MS. Racial variations in the incidence of trisomy 21. Br J Obstet Gynaecol 1986;93:587-99.
5. Bogart MH, Jones OH, Felder RA, Best RG, Bradley I, Butts W, et al. Prospective evaluation of maternal serum human chorionic gonadotrophin levels in 3428 pregnancies. Am J Obstet Gynaecol 1991;165:663-7. [Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    StumbleUpon    Technorati    What's this?

Relevant Articles

Assay used in serum screening for Down's syndrome may depend on ethnic make up of population

L Gilbert, Jon Nicholl, and Anthony Mander
BMJ 1996 313: 297. [Extract] [Full Text]

Ethnic differences in outcome of serum screening for Down's syndrome

Clare Ford, Alison Moore, William A Bartlett, and Alan F Jones
BMJ 1996 312: 1040-1041. [Extract] [Full Text]

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

P J Holloway and S Bulusu
BMJ 1996 312: 1041. [Extract] [Full Text]

Median values and maternal weight specific to the ethnic group should be used

Kevin Spencer
BMJ 1996 312: 1041. [Extract] [Full Text]

This article has been cited by other articles:

• Dormandy, E., Michie, S., Hooper, R., Marteau, T. M (2005). Low uptake of prenatal screening for Down syndrome in minority ethnic groups and socially deprived groups: a reflection of women's attitudes or a failure to facilitate informed choices?. Int J Epidemiol 34: 346-352 [Abstract] [Full text]  
• Gilbert, L, Nicholl, J., Mander, A. (1996). Assay used in serum screening for Down's syndrome may depend on ethnic make up of population. BMJ 313: 297b-297 [Full text]  
• Holloway, P J, Bulusu, S (1996). For appreciable ethnic differences in false positive rates, use different cut off points. BMJ 312: 1041-1041 [Full text]  
• Spencer, K. (1996). Median values and maternal weight specific to the ethnic group should be used. BMJ 312: 1041a-1041 [Full text]