Values in parentheses are 95% confidence intervals. Hest value = 2.37; 902 degrees of freedom; P < 0.02.

Plasmid analysis was used in the investigation of an outbreak of nosocomial Legionnaires' disease in four patients. Serogroup 1 strains were isolated from two patients, the air-conditioning cooling tower, and two hot-water tanks. All serogroup 1 strains contained two plasmids with approximate molecular masses of 21 and 48 megadaltons (Mdal). The serogroup 1 strain found in the cooling-tower isolate also contained an additional 1.9 Mdal-plasmid. Restriction-endonuclease analysis of the 21-Mdal plasmid that was present in patient and hot water-tank isolates revealed identical EcoRI and HaeIII fragment patterns. Digestion of the similarly sized plasmid in the cooling-tower isolate resulted in a unique fragment pattern. The data provide direct bacteriologic evidence implicating the hot-water tanks rather than the cooling tower as the source of the infecting strain.

No abstract

A cohort of 570 untreated pregnant women from Burkina Faso was studied to assess the influence of epidemiologic factors on malaria infection, which was quantified as the mean of serial, season-adjusted parasitemia measuremects (mean parasite density [MPD]) carried out during the last five months of gestation. A significant effect of the area of maternal residence on the MPD was found ( P < 0.003) and was probably. due to geographic differences in mosquito transmission conditions. The strong relationship observed between parity and malaria infection ( P < O.OOOl), with MPD levels decreasing as the number of gestations increased, confirms that primigravidae are a high-risk group whose protection should be a priority. After adjustment for two relevant epidemiologic factors (i.e., area of residence and parity), the residual MPD values fitted a mixture of two distributions. This result supports the view that a major gene is involved in the determination of malaria infection intensities and is consistent with the results of a recent familial study in Cameroon.

The environmental and socioeconomic risk factors for preterm delivery were assessed in a West African urban population (Bobo-Dioulasso, Burkina Faso). The study population were 102 cases of preterm delivery matched with 102 controls obtained from 4124 sequential deliveries which occurred between May and October 1989 in the three maternity centres in the city. The univariate analysis identified the risk factors as age (< 20 years), primiparity, marital status (single), low frequency of antenatal visits, death of a previous child and level of education of the mother. The following risk factors identified by multivariate analysis (logistic regression) are consistent with those identified in previous studies: youth of the mother, primiparity (P = 0.01) and death of a previous child (P < 0.05). On the other hand, in this study, the level of education of the parent was identified as an independent risk factor (P < 0.001). This finding could be used to determine a target population for prevention programmes.