Objective To analyse associations of ethnicity and consanguinity with causes of infant death.

Design Retrospective population based study of all infant deaths in a health district in a year.

Methods We obtained a list of all infant deaths from PCT and reviewed notes on each death.

Results 136 infants died during 2006–2007 (IMR: 8.2)1; 47 were White British and 73 Pakistani. Consanguinity was known to be present in 55% of Pakistani families, and in no White British families.

Congenital anomaly (CA) as a cause of death was significantly more among Pakistani than White British families (13 of 47, 27.7%, vs 41 of 73, 56.2%; p<0.02). Other causes of death did not differ by ethnicity. Recessive conditions accounted for 13 of 41 (32%) CA deaths in Pakistani population and none in White British.

Univariate analysis showed consanguinity and ethnicity were significantly associated with death from CA (p<0.001). Multivariate analysis showed that only consanguinity remained associated with CA as a cause of neonatal deaths.

Clinician review of cause of death indicated 54 deaths were due to CA. Of these, ONS grouped only 41 as being deaths from CA.

Conclusion Excess Pakistani infant mortality2 in this population appears to be partly explained by the greater prevalence of consanguinity. ONS cause of death grouping may underreport the impact of rare pathologies on deaths from CA in this ethnic group.