Few studies have explored diagnosis delay by tuberculosis (TB) patients and its effects on the rate of infection among their close contacts. A cross-sectional study of the close contacts of 505 newly diagnosed TB patients was conducted in a TB referral centre in Sana'a, Yemen from 2008 to 2010. Only the close contacts of 89 new TB patients agreed to participate and completed the tuberculin skin test (TST). Of the 239 close contacts investigated, 133 (55.6%) had a positive TST result. Index patients were classified as long or short diagnosis delay (above or below the median). There was no significant difference in the number of infected close contacts between long and short delay index patients (Mann-Whitney U-test). A larger sample size, with more incentives for patients to participate and the use of other investigative tools could provide a better picture of the pattern of TB transmission among all contacts.