Demographic data for HCWs were provided by the Human Resources Department of the hospital. Population in December was taken as the study population for each year. Incidence rates of TB for Turkey's general population were taken from the annual statistics report of the Ministry of Health [7].

Twenty three of the cases were pulmonary TB and 8 were extra-pulmonary. In the pulmonary TB patients, chest x-ray revealed typical lung parenchyma pathology. There were no cases of active TB in the families of HCWs.

Mean incidence of TB for 1991 to 2000 was 96 per 100,000 for all HCWs (relative risk: 2.71), 79 per 100,000 for doctors (relative risk: 2.2), 14 per 100,000 for nurses and 121 per 100,000 (relative risk: 3.4) for other professionals. Mean incidence of TB for Turkey as a whole was 35.4 per 100,000 [7]. Fig. ​Fig.11 shows the incidence of TB in our study and in the general population.

The demographic characteristics of TB cases among our hospital HCWs were different from the general population. In Turkey, it has been shown that TB is more common among persons with a mean age of 37.5 years (range 20–39) and 85% of patients are male [9]. However, in our hospital staff, the mean age of TB patients was 27.7 years and 50 % of them were female. These differences support the idea that TB in HCWs was of nosocomial origin.

There are two limitations in our study. First, we do not have information about staff that left the hospital to pursue another occupation. Our data was limited to the time that they worked in the hospital. The second limitation concerns the incidence rates of TB in the general population. We accepted the Ministry of Health figures but we estimate that the reported incidence rates are less than the true rates due to the deficiency in registering TB cases in Turkey.

Those that have investigated TB infection risk by tuberculin testing have shown that HCWs have a higher infection risk than the normal population [1-3]. In Turkey, BCG vaccination is routinely given twice in a person's life; at birth and at 7 years of age. As a result, tuberculin skin test surveys are not relevant for Turkey. All of the HCWs in our study had been vaccinated at least once.

Previous studies from other countries have demonstrated different results for TB in HCWs [4-6,8,10-16]. Kruuer et al. showed that TB risk in Estonia was higher for HCWs than the rest of population [4]. Similar results have been found in Serbia [5], Malawi [6], Japan [10], North America and Western Europe [12-14]. However, two studies had different findings [15,16]. Firstly, in the UK, McKenna showed that risk of TB among HCWs was similar to that of the general population. Secondly, Raitio found that the overall risk of TB among HCWs was lower than in the general population. The incidence of TB in Turkey in 2000 of 26.3 per 100,000 is lower than the rates in Estonia, Malawi and Serbia but higher than North America and Western Europe. Nevertheless, our results are in line with those of other countries.

In Turkey there is no specific TB prevention program for HCWs. There are several chest clinics where HCWs and TB patients use surgical masks but there is only one clinic which has negative ventilation for its TB ward. In North America, on the other hand, TB prevention programs for HCWs have been established and performance is regularly controlled [17]. In spite of this structural difference, the similar, higher TB risk for HCWs in Turkey and North America is significant. In our Department of Chest Diseases, the last TB case in an HCW was in 1991. Surgical masks have been used in the department since 1992. A separate TB ward was set up in 1996 and in 1998 it was fitted with a central negative ventilation system. We observed that in other hospital departments, with the exception of nurses, staff did not take precautions against TB. The incidence of TB among nurses was very low in comparison with other staff. All the nurses in our hospital have a regular post-graduate education program but the doctors and other staff do not. We speculate that the low incidence of TB in nurses is dependent on this regular education. TB is still a disease that affects low income groups in Turkey. The level of income for HCWs is higher than the poverty threshold. Indeed, the incidence of TB in that income group would have been expected to be lower than in the general population.

Unfortunately, we do not have detailed information about TB in Turkey that takes account of different income levels. However, we predict that TB incidence in nurses in our hospital will be similar to non-medical personnel who have the same income as HCWs. Two other factors may account for the different TB risk in nurses compared with other health professionals. Nurses usually work in wards where it is known if a patient has TB, therefore, nurses can take steps to protect themselves against infection. Doctors and other health professionals work in out-patient and several other departments at the same time. In these circumstances there may be patients with undiagnosed TB, leading to increased risk of infection.

All authors read and approved the final manuscript.