The pathology and pathogenesis of experimental severe acute respiratory syndrome and influenza in animal models.

Respiratory viruses that emerge in the human population may cause high morbidity and mortality, as well as concern about pandemic spread. Examples are severe acute respiratory syndrome coronavirus (SARS-CoV) and novel variants of influenza A virus, such as H5N1 and pandemic H1N1. Different animal models are used to develop therapeutic and preventive measures against such viruses, but it is not clear which are most suitable. Therefore, this review compares animal models of SARS and influenza, with an emphasis on non-human primates, ferrets and cats. Firstly, the pathology and pathogenesis of SARS and influenza are compared. Both diseases are similar in that they affect mainly the respiratory tract and cause inflammation and necrosis centred on the pulmonary alveoli and bronchioles. Important differences are the presence of multinucleated giant cells and intra-alveolar fibrosis in SARS and more fulminant necrotizing and haemorrhagic pneumonia in H5N1 influenza. Secondly, the pathology and pathogenesis of SARS and influenza in man and experimental animals are compared. Host species, host age, route of inoculation, location of sampling and timing of sampling are important to design an animal model that most closely mimics human disease. The design of appropriate animal models requires an accurate pathological description of human cases, as well as a good understanding of the effect of experimental variables on disease outcome.

Modification of the ferret model for pneumonia from seasonal human influenza A virus infection.

The primary complication of seasonal influenza in humans is viral pneumonia. A conventional animal model--intranasal inoculation of ferrets with 10(6) median tissue culture infectious dose of virus--results in disease that is neither consistent nor comparable with severe viral pneumonia in humans. Therefore, the authors modified the experimental procedures by increasing the median tissue culture infectious dose to 10(9) and by inoculating via the intratracheal route, testing these procedures with H1N1 strains (A/Bilthoven/3075/1978 and A/Netherlands/26/2007) and H3N2 strains (A/Bilthoven/16190/1968 and A/Netherlands/177/2008) of seasonal influenza virus. The ferrets of all groups (n = 3 per virus strain) had clinical signs, increased body temperature, virus excretion from day 1, loss of body weight, and increased relative lung weight at 4 days postinoculation. All ferrets had severe pulmonary consolidation, and histologic examination revealed moderate to severe necrotizing bronchointerstitial pneumonia with severe edema, necrosis of alveolar epithelium, inflammatory infiltrates in alveolar septa and lumina, epithelial regeneration, and perivascular and peribronchiolar inflammatory infiltrates. The lesions were associated with the presence of influenza virus antigen in respiratory epithelium by immunohistochemistry. Although all 4 virus strains caused pulmonary lesions of comparable severity, virus isolation in the lungs, trachea, nasal concha, and tonsils showed higher mean virus titers in the H1/07 and H3/68 groups than in the H1/78 and H3/08 groups. In conclusion, the above H1N1 and H3N2 strains cause severe pneumonia in ferrets by use of the modified experimental procedures and provide a good model for pneumonia caused by seasonal influenza A virus infection in humans.

Pandemic 2009 H1N1 influenza virus causes diffuse alveolar damage in cynomolgus macaques.

The pathogenesis of lower respiratory tract disease from the pandemic 2009 H1N1 (H1N1v) influenza A virus is poorly understood. Therefore, either H1N1v virus or a seasonal human H1N1 influenza A virus was inoculated into cynomolgus macaques as a nonhuman primate model of influenza pneumonia, and virological, pathological, and microarray analyses were performed. Macaques in the H1N1v group had virus-associated diffuse alveolar damage involving both type I and type II alveolar epithelial cells and affecting an average of 16% of the lung area. In comparison, macaques in the seasonal H1N1 group had milder pulmonary lesions. H1N1v virus tended to be reisolated from more locations in the respiratory tract and at higher titers than seasonal H1N1 virus. In contrast, differential expression of messenger RNA transcripts between H1N1v and seasonal H1N1 groups did not show significant differences. The most upregulated genes in H1N1v lung samples with lesions belonged to the innate immune response and proinflammatory pathways and correlated with histopathological results. Our results demonstrate that the H1N1v virus infects alveolar epithelial cells and causes diffuse alveolar damage in a nonhuman primate model. Its higher pathogenicity compared with a seasonal H1N1 virus may be explained in part by higher replication in the lower respiratory tract.

A single immunization with CoVaccine HT-adjuvanted H5N1 influenza virus vaccine induces protective cellular and humoral immune responses in ferrets.

Highly pathogenic avian influenza A viruses of the H5N1 subtype continue to circulate in poultry, and zoonotic transmissions are reported frequently. Since a pandemic caused by these highly pathogenic viruses is still feared, there is interest in the development of influenza A/H5N1 virus vaccines that can protect humans against infection, preferably after a single vaccination with a low dose of antigen. Here we describe the induction of humoral and cellular immune responses in ferrets after vaccination with a cell culture-derived whole inactivated influenza A virus vaccine in combination with the novel adjuvant CoVaccine HT. The addition of CoVaccine HT to the influenza A virus vaccine increased antibody responses to homologous and heterologous influenza A/H5N1 viruses and increased virus-specific cell-mediated immune responses. Ferrets vaccinated once with a whole-virus equivalent of 3.8 microg hemagglutinin (HA) and CoVaccine HT were protected against homologous challenge infection with influenza virus A/VN/1194/04. Furthermore, ferrets vaccinated once with the same vaccine/adjuvant combination were partially protected against infection with a heterologous virus derived from clade 2.1 of H5N1 influenza viruses. Thus, the use of the novel adjuvant CoVaccine HT with cell culture-derived inactivated influenza A/H5N1 virus antigen is a promising and dose-sparing vaccine approach warranting further clinical evaluation.

Preclinical evaluation of a modified vaccinia virus Ankara (MVA)-based vaccine against influenza A/H5N1 viruses.

Highly pathogenic avian influenza viruses of the H5N1 subtype are responsible for an increasing number of infections in humans since 2003. More than 60% of the infections is lethal and new infections are reported frequently. In the light of the pandemic threat caused by these events the rapid availability of safe and effective vaccines is desirable. Modified vaccinia virus Ankara (MVA) expressing the HA gene of an influenza A/H5N1 virus is a promising candidate vaccine that induced protective immunity against infection with homologous and heterologous influenza A/H5N1 viruses in mice. We also evaluated the recombinant MVA vector expressing the HA of influenza A/H5N1 virus A/Vietnam/1194/04 (MVA-HA-VN/04) in non-human primates. Cynomolgus macaques were immunized twice and then challenged with influenza virus A/Vietnam/1194/04 (clade 1) or A/Indonesia/5/05 (clade 2.1) to assess the level of protective immunity. Immunization with MVA-HA-VN/04 induced (cross-reactive) antibodies and prevented virus replication in the upper and lower respiratory tract and the development of severe necrotizing bronchointerstitial pneumonia. Therefore MVA-HA-VN/04 is a promising vaccine candidate for the induction of protective immunity against highly pathogenic avian influenza A/H5N1 viruses.

Infection of mice with a human influenza A/H3N2 virus induces protective immunity against lethal infection with influenza A/H5N1 virus.

The transmission of highly pathogenic avian influenza (HPAI) A viruses of the H5N1 subtype from poultry to man and the high case fatality rate fuels the fear for a pandemic outbreak caused by these viruses. However, prior infections with seasonal influenza A/H1N1 and A/H3N2 viruses induce heterosubtypic immunity that could afford a certain degree of protection against infection with the HPAI A/H5N1 viruses, which are distantly related to the human influenza A viruses. To assess the protective efficacy of such heterosubtypic immunity mice were infected with human influenza virus A/Hong Kong/2/68 (H3N2) 4 weeks prior to a lethal infection with HPAI virus A/Indonesia/5/05 (H5N1). Prior infection with influenza virus A/Hong Kong/2/68 reduced clinical signs, body weight loss, mortality and virus replication in the lungs as compared to naive mice infected with HPAI virus A/Indonesia/5/05. Priming by infection with respiratory syncytial virus, a non-related virus did not have a beneficial effect on the outcome of A/H5N1 infections, indicating that adaptive immune responses were responsible for the protective effect. In mice primed by infection with influenza A/H3N2 virus cytotoxic T lymphocytes (CTL) specific for NP(366-374) epitope ASNENMDAM and PA(224-232) SCLENFRAYV were observed. A small proportion of these CTL was cross-reactive with the peptide variant derived from the influenza A/H5N1 virus (ASNENMEVM and SSLENFRAYV respectively) and upon challenge infection with the influenza A/H5N1 virus cross-reactive CTL were selectively expanded. These CTL, in addition to those directed to conserved epitopes, shared by the influenza A/H3N2 and A/H5N1 viruses, most likely contributed to accelerated clearance of the influenza A/H5N1 virus infection. Although also other arms of the adaptive immune response may contribute to heterosubtypic immunity, the induction of virus-specific CTL may be an attractive target for development of broad protective vaccines. Furthermore the existence of pre-existing heterosubtypic immunity may dampen the impact a future influenza pandemic may have.

Pathology and virus distribution in chickens naturally infected with highly pathogenic avian influenza A virus (H7N7) During the 2003 outbreak in The Netherlands.

The largest recorded outbreak of highly pathogenic avian influenza virus of the subtype H7N7 occurred in The Netherlands in 2003. We describe the immunohistochemical and histopathologic findings of 3 chickens naturally infected during this outbreak. Influenza virus antigen occurred in endothelial cells and mononuclear cells of all tissues examined and occurred in parenchymal cells of heart, lung, kidney, pancreas, and trachea, often associated with multifocal inflammation and necrosis. These findings are consistent with the acute stage of highly pathogenic avian influenza from other subtypes. In the severely edematous wattle skin, most endothelial cells contained virus antigen, while in all other tissues virus antigen was only detected in a few endothelial cells. Virus histochemistry showed that this H7N7 virus attached to more endothelial cells in wattle skin than in other vascular beds. This might explain, at least partly, the tropism of the virus and the associated severity of lesions in this tissue.

Recombinant modified vaccinia virus Ankara expressing the hemagglutinin gene confers protection against homologous and heterologous H5N1 influenza virus infections in macaques.

BACKGROUND: Highly pathogenic avian influenza viruses of the H5N1 subtype have been responsible for an increasing number of infections in humans since 2003. More than 60% of infected individuals die, and new infections are reported frequently. In light of the pandemic threat caused by these events, the rapid availability of safe and effective vaccines is desirable. Modified vaccinia virus Ankara (MVA) expressing the hemagglutinin (HA) gene of H5N1 viruses is a promising candidate vaccine that induced protective immunity against infection with homologous and heterologous H5N1 influenza virus in mice. METHODS: In the present study, we evaluated a recombinant MVA vector expressing the HA gene of H5N1 influenza virus A/Vietnam/1194/04 (MVA-HA-VN/04) in nonhuman primates. Cynomolgus macaques were immunized twice and then were challenged with influenza virus A/Vietnam/1194/04 (clade 1) or A/Indonesia/5/05 (clade 2.1) to assess the level of protective immunity. RESULTS: Immunization with MVA-HA-VN/04 induced (cross-reactive) antibodies and prevented virus replication in the upper and lower respiratory tract and the development of severe necrotizing bronchointerstitial pneumonia. CONCLUSION: Therefore, MVA-HA-VN/04 is a promising vaccine candidate for the induction of protective immunity against highly pathogenic H5N1 avian influenza viruses in humans.

Pathology of experimental SARS coronavirus infection in cats and ferrets.

The pathology of severe acute respiratory syndrome-coronavirus (SARS-CoV) infection in cats and ferrets is poorly described, and the distribution of angiotensin-converting enzyme 2 (ACE2), a receptor for SARS-CoV, in the respiratory tracts of these species is unknown. We observed SARS-CoV antigen expression and lesions in the respiratory tracts of 4 cats and 4 ferrets at 4 days postinoculation and ACE2 expression in the respiratory tracts of 3 cats and 3 ferrets without infection. All infected cats and ferrets had diffuse alveolar damage associated with SARS-CoV antigen expression. A novel SARS-CoV-associated lesion was tracheo-bronchoadenitis in cats. SARS-CoV antigen expression occurred mainly in type I and II pneumocytes and serous cells of tracheo-bronchial submucosal glands of cats and in type II pneumocytes of ferrets. ACE2 expression occurred mainly in type I and II pneumocytes, tracheo-bronchial goblet cells, serous epithelial cells of tracheo-bronchial submucosal glands in cats, and type II pneumocytes and serous epithelial cells of tracheo-bronchial submucosal glands in ferrets. In conclusion, the pathology of SARS-CoV infection in cats and ferrets resembles that in humans except that syncytia and hyaline membranes were not observed. The identification of tracheo-bronchoadenitis in cats has potential implications for SARS pathogenesis and SARS-CoV excretion. Finally, these results show the importance of ACE2 expression for SARS-CoV infection in vivo: whereas ACE2 expression in type I and II pneumocytes in cats corresponded to SARS-CoV antigen expression in both cell types, expression of both ACE2 and SARS-CoV antigen in ferrets was limited mainly to type II pneumocytes.

A clinicopathological study of 52 feline epulides.

A retrospective study was performed to characterize 52 new cases of feline epulides between 1995 and 2001, with clinical and pathological results classified according to Head's histopathologic criteria for canine epulides. The incidence of the fibromatous, acanthomatous, ossifying, and giant cell epulis were respectively 57.7% (30/52), 7.7% (4/52), 5.8% (3/52), and 28.8% (15/52). Giant cell epulides presented significant differences in clinical behavior compared with the fibromatous type, including rapid growth (P < .0001), presence of ulcerative changes (P < .01), and rapid recurrence after surgery (P < .01) from which euthanasia was judged necessary in 4 cases. Fifteen giant cell epulides were additionally examined in order to characterize the lesion both histochemically and immunohistochemically and to investigate the origin of the multinucleated giant cells (MGCs). Van Gieson staining showed osteoid and woven bone formation in 11 cases. Both the MGCs and a fraction of the mononuclear cells were positive for vimentin, tartrate-resistant acid phosphatase (TRAP), a commonly accepted marker for osteoclasts, and the polyclonal antibody receptor activator of nuclear factor kappabeta (RANK), a cytokine leading to the differentiation of osteoclast progenitors into mature osteoclasts in presence of its ligand. MGCs were negative for smooth muscle actin, MIB-1, and factor VIII. The giant cell epulis may be a variant of the fibromatous and ossifying epulis in which extensive ulceration and inflammation results in increased osteoclastic activity. The osteoclast-like giant cells are most likely formed from a monocyte/macrophage-like osteoclast precursor that differentiates into osteoclasts under the influence of mononuclear osteoblast-like stromal cells.

Renal cysts, nephrolithiasis, and ovarian cysts in a donkey suspected of having been sexually abused.

A female donkey was thought by its owner to have been sexually abused because it had severe perineal swelling. Besides the perineal swelling and a very small vaginal erosion, there were no other abnormalities at clinical examination. Haematology and blood biochemistry revealed an increased leukocyte count, an elevated blood lactate concentration, and a low ionized calcium concentration. During night the donkey's condition deteriorated and it was euthanized in the morning. At necropsy severe haemorrhages were found within the subserosa of the caudal abdomen. Both kidneys were polycystic, and multiple calculi were found in the right kidney. Both ovaries had multiple cysts. Lesions (fibrosis and mineralization) were found in the liver, lungs, and mesenteric artery and were suggestive of an earlier parasitic infection. There was no evidence of sexual abuse.

Kinetics and organ distribution of allogeneic natural killer lymphocytes transfused into patients suffering from renal cell carcinoma.

The transfusion of natural killer (NK) lymphocytes into patients suffering from malignant diseases is an approach of current interest in the field of immunotherapy. Little is known about the organ distribution, survival, and clearance of donor immune effector cells in cellular therapy, and no reports exist on these important parameters considering NK cells in particular or any other type of allogeneic lymphocytes in humans. In the context of a clinical Phase I/II study we examined the distribution of transfused allogeneic NK cells in patients suffering from renal cell carcinoma. The NK cells were ex vivo cultivated and activated before transfusion. To assess the circulation of the transfused cells in the peripheral blood, we used a nested PCR technique to detect HLA DRB1 alleles of the NK cell donors. Post-transfusion, all patients showed evidence of circulating donor cells for up to 3 days. After 7 days, all donor cells were cleared from the blood to undetectable levels. To assess organ distribution, (111)In-labeled NK cells were injected and monitored by whole-body scintiscans. A distribution to the whole body, with preference for liver, spleen, and bone marrow, was observed after a short initial uptake in the lungs. No activity was observed in lymphatic nodes. A total of 2/4 evaluable metastases showed a clear accumulation of transfused NK cells. The half-life corrected activity in all body compartments remained almost constant over the 6-day observation period in concordance with the absence of any excretion of radioactivity. This may indicate an extended survival of the transfused cells, despite their foreign nature, in the host organism.

Upregulation of IFN-gamma and soluble interleukin-2 receptor release and altered serum cortisol and prolactin concentration during general anesthesia.

The effects of surgery, surgical stress, and anesthesia compromise the optimal function of the immune system. Recent studies demonstrate the influence of anesthesia on the immune response by modulation of neural-immune interactions. To evaluate the immunologic effects of general anesthesia with the hypnotic agent propofol and the opioid fentanyl, two drugs used frequently in anesthesia, we studied 30 patients undergoing elective orthopedic surgery before and during narcosis. We found a significant enhancement of interferon-gamma (IFN-gamma) and soluble interleukin-2 receptor (sIL-2R) release in lipopolysaccharide (LPS)-stimulated whole blood cultures after induction of anesthesia. Similar results were observed in cultures stimulated with polyclonal T cell activators, such as staphylococcal enterotoxin B (SEB) and phytohemagglutinin (PHA). IL-1beta and IL-8 release was not affected, but the anti-inflammatory cytokine IL-10 decreased after skin incision. Serum prolactin significantly increased immediately after induction of anesthesia, whereas serum cortisol levels declined. Our results point to enhanced proinflammatory T lymphocyte and natural killer (NK) cell activity, probably caused by prolactin and cortisol modulation in the serum. This may disturb the balance of human proinflammatory and anti-inflammatory pathways during surgery and general anesthesia.

Extracellular polysaccharides of modified strains of Erwinia spp.

The structure of the extracellular polysaccharide (EPS) produced by Erwinia chrysanthemi strain A2148 has been determined using low pressure size-exclusion and anion-exchange chromatographies, high pH anion-exchange chromatography, glycosyl-linkage analysis, and 1D 1H NMR spectroscopy. The polysaccharide is structurally similar, if not identical, to the EPS produced by E. chrysanthemi strain A350. A streptomycin-resistant strain of E. chrysanthemi Ech6 (Ech6S(+)) has been generated and has an elevated production of EPS, as does a streptomycin-resistant strain (Ech9Sm6) of E. chrysanthemi Ech9. These modified E. chrysanthemi spp. have been ribotyped and found to be closely related to their parent strains.

[Mechanical autotransfusion procedures. The effect cytokines and leukocytes on washed erythrocyte concentrate]. / Maschinelle Autotransfusionsverfahren. Anwendung und Einfluss auf Zytokine und Leukozyten im autologen gewaschenen Erythrozytenkonzentrat.

BACKGROUND: The use of mechanical autotransfusion devices has been sporadically associated with adverse effects including pulmonary dysfunction and systemic inflammatory response. Stimulated immune cells and proinflammatory cytokines are suspected mediators of these complications. This study was designed to evaluate whether mechanical autotransfusion stimulates immune cells in wound blood thus leading to an increase in cytokines. METHODS: The wound blood of 100 patients undergoing total knee arthroplasty was collected and processed using four different devices in a randomized order. Leucocyte and cytokine concentrations (TNF-alpha, IL-6, IL-10) were determined in the collected wound blood and in the washed erythrocyte concentrate. RESULTS: IL-6 concentrations in the collected wound blood were markedly higher. TNF-alpha and IL-10 could also be detected at lower concentrations. Cell washing reduced IL-6 and IL-10 levels efficiently, while TNF-alpha concentrations were significantly increased. Leucocyte concentrations were decreased only slightly. A content of 1.0 x 10(9) leucocytes remained in the processed blood. CONCLUSION: High concentrations of IL-6 in the collected wound blood are reduced extensively by all mechanical autotransfusion devices investigated. However, a large amount of leucocytes remains in the processed blood. Considering the increase in TNF-alpha concentration after the washing, an elevated activity of mononuclear cells in the blood product must be assumed.

[Cytokine release and changes in mononuclear cells in peripheral blood under the influence of general anesthesia]. / Zytokinfreisetzung und Verteilung mononukleärer Zellen im peripheren Blut unter dem Einfluss der Allgemeinnarkose.

OBJECTIVE: Anaesthetic agents are believed to have an adverse effect on human immune defense mechanisms. We investigated changes in peripheral immune cell numbers such as natural killer (NK) cells, B cells and T lymphocyte subpopulations (CD4+ and CD8+ cells) and differences in cytokine production after stimulation with different mitogens before and during narcosis. METHODS: We studied 30 patients undergoing elective orthopedic surgery. Stimulatory experiments were performed with the mitogens lipopolysaccharide (LPS), phytohemagglutinin (PHA) and inactivated Newcastle Disease Virus (NDV). RESULTS: During general anaesthesia with fentanyl, thiopental, isoflurane and nitrous oxide there was a significant decrease of circulating NK cells in the peripheral blood accompanied by a significant increase of B cells and CD8+ T lymphocytes. We detected a significant anesthesia-associated increase of interferon (IFN)-gamma, INF-alpha, tumor necrosis factor (TNF)-alpha and soluble interleukin-2 receptor (sIL-2R) synthesis after stimulation with different mitogens while interleukin (IL)-1 beta and IL-6 protein did not change significantly. After the beginning of surgery CD8-positive cells showed a return to control values and NK cell number increased slightly. CONCLUSION: These findings suggest that general anaesthesia interferes with immune cell number and immune cell response. This may explain the clinically well-recognized disturbance of human immunity after surgery and general anaesthesia.

The effects of general anesthesia on human peripheral immune cell distribution and cytokine production.

Anesthetic agents are believed to have an adverse effect on human immune defense mechanisms. We investigated changes in peripheral immune cell numbers such as natural killer (NK) cells, B cells, and T lymphocyte subpopulations (CD4+ and CD8+ cells) and differences in cytokine production after stimulation with different mitogens before and during narcosis. We studied 30 patients undergoing elective orthopedic surgery. Stimulatory experiments were performed with the mitogens lipopolysaccharide, phytohemagglutinin A, and inactivated Newcastle disease virus. During general anesthesia with fentanyl, thiopental, and isoflurane, there was a significant decrease of circulating NK cells in the peripheral blood accompanied by a significant increase of B cells and CD8+ T lymphocytes. We detected a significant anesthesia-associated increase of interferon (IFN)-gamma, IFN-alpha, tumor necrosis factor-alpha, and soluble interleukin-2 receptor (sIL-2R) synthesis after stimulation with different mitogens while interleukin (IL)-1 beta and IL-6 protein did not change significantly. After the beginning of surgery, CD8-positive cells showed a return to control values and NK cell number increased slightly. These findings suggest that general anesthesia interferes with immune cell number and immune cell response. This may explain the clinically well-recognized disturbance of human immunity after surgery and general anesthesia.

Stimulation of whole blood cultures in patients with ankylosing spondylitis by a mitogen derived from Mycoplasma arthritidis (MAS) and other mitogens.

In this study we compared cytokine production and cell proliferation of immunocompetent cells derived from patients with ankylosing spondylitis (AS) to those from healthy blood donors using a whole blood assay. To this end, blood cell cultures were stimulated with the superantigens MAS (Mycoplasma arthritidis supernatant) and staphylococcal enterotoxin B (SEB) and the plant lectins phytohaemagglutinin (PHA) and concanavalin A (Con A). The number of white blood cells (WBC) and lymphocyte subsets were also determined. Cell proliferation and levels of interferon-gamma (IFN-gamma), interleukin-1 beta (IL-1 beta) and interleukin-6 (IL-6) were measured after stimulation with the different mitogens. An ELISA test was used to analyse supernatant cytokine levels. Individuals with AS showed significantly lower IFN-gamma concentrations and markedly lower cell proliferation rates with all tested mitogens than healthy controls, while there was no significant difference in IL-6 synthesis. IL-1 beta levels were slightly impaired in the patient group, but only blood cell cultures stimulates with MAS showed a statistical significance. Furthermore, there was a significant elevation of leucocytes and lymphocytes in patients with AS resulting in higher numbers of CD4-positive cells, which implies a higher CD4:CD8 cell ratio. CD19- and CD8-positive cells were not significantly distinct compared to healthy controls. This deviation in cytokine levels and cell proliferation points to a suppression of T lymphocytes. A disturbed T-lymphocyte function may play a part in the pathogenesis of AS.

Use of the intramuscular relative-dose-response test to predict bronchopulmonary dysplasia in premature infants.

We tested the hypothesis that an intramuscular relative dose response (IM-RDR) on day 1 of life would more accurately predict which premature infants would develop bronchopulmonary dysplasia (BPD) than single measurements of retinol, retinyl palmitate (RP), or retinol binding protein (RBP). Seventy-five premature infants < or = 32 wk gestation had the IM-RDR on day 1 of life. An RDR > or = 25% occurred in 6 of 37 infants who did not develop BPD compared with 15 of 38 infants who developed BPD (P = 0.025). Retinol, RP, and RBP on day 1 were not different between the groups. BPD infants who received postnatal dexamethasone during their hospital course had a higher day-28 baseline retinol concentration (1.19 +/- 0.15 mumol/L) than did the group with no BPD (0.82 +/- 0.06 mumol/L) (P = 0.03). However, the effect of postnatal dexamethasone on serum retinol was biphasic, rising initially, and then declining after 8-12 d. RP values at time 0 and 5 h on day 28 were higher than day 1 values in both infants without BPD and infants with BPD who did not receive dexamethasone. Retrospective analysis also revealed a significant correlation between a day-1 RDR > or = 25% and the incidence of intraventricular hemorrhage in these premature infants. Because the IM-RDR is more helpful in predicting the development of BPD than single serum retinol and RP analyses, this test could be useful in determining which premature infants might benefit from supplemental vitamin A for BPD.