We thank Ozdemir et al. for their comments and suggestions pertaining to our recent article entitled “Necrotizing enterocolitis is associated with Ureaplasma colonization in preterm infants.” The authors note that they have unpublished data based upon the evaluation of 224 infants <34 wk gestation for Ureaplasma colonization by one or more tracheal or nasopharyngeal aspirate cultures. The authors did not find a significant association between Ureaplasma colonization and the development of necrotizing enterocolitis (NEC) in their cohort but other factors such as red blood cell transfusions (85 ± 45.9 versus 50.8 ± 32.3; p < 0.01) and proven late-onset sepsis (61.3% versus 37.8%; p = 0.01) were significant. Similar to our findings, cord serum IL-6 levels were higher in Ureaplasma-colonized NEC infants. Ozdemir et al. further stated that after multivariate analysis, only the number of red blood cell transfusions received remained significantly associated with the development of NEC (OR, 1.4; 95% CI, 1.18–1.74; p = 0.01).

The pathogenesis of NEC remains poorly understood. Although prematurity is the greatest risk factor for developing NEC, epidemiological studies suggest a multifactorial etiology (1,2). Intestinal immaturity heralded by abnormal microbial intestinal colonization, alterations in intestinal blood flow, and a highly immunoreactive intestinal mucosa are important predisposing factors to the extensive mucosal damage seen in NEC. Contrary to our findings, Ozdemir et al. did not find an association between Ureaplasma colonization and NEC stage ≥2 in their cohort. Several factors may explain the differences in these findings. Our sample size was larger than both the authors' and the Perzigian cohorts (3). We analyzed data from 368 subjects, and NEC incidence remained greater in Ureaplasma colonized compared with negative infants even when substratified to only include infants ≤28 wk gestation at birth. Because culture methodology details were not provided by Ozdemir et al., we cannot comment on whether differences in Ureaplasma detection may have contributed to differences in study outcomes. Although the majority of subjects in our sample were of African-American descent, it is not clear whether the authors included race in their descriptive variables. The incidence of NEC is known to be higher in the African-American population in the United States which is attributed to the higher rate of preterm delivery in African-Americans (4).

Several retrospective studies including the one by the authors have shown a possible association between the timing of packed red blood cell transfusions and the subsequent development of NEC (5). Other recent reports do not define a clear relationship (6). It is therefore difficult to conclusively state causality. Transfusion of the anemic preterm infant has been suggested to cause a reperfusion-type injury to an already compromised immature intestinal mucosa (7). Ingestion of Ureaplasma-infected amniotic fluid by the fetus could potentially initiate an excessive systemic and local intestinal inflammatory response and thus predispose these very LBW infants to NEC. Anemia-related tissue hypoxia may also play a role (8). Although we appreciate the comments of Ozdemir et al., the results from our predominantly African-American sample indicates that a reevaluation of our data along their suggested lines would be of no additional value to our study.