Probiotic use in pediatric kidney transplant recipients: What are current practices, and are they evidence-based? A pediatric nephrology research consortium study.

BACKGROUND: Probiotics are living microorganisms that may confer health benefits to their host if administered in sufficient quantities. However, data on the use of probiotics in transplant recipients are scarce. METHOD: This multi-center survey of pediatric nephrologists aimed to examine variations in practice regarding the use of probiotics in pediatric kidney transplant recipients. The survey was conducted via a 10-item questionnaire (developed in Survey Monkey) administered to pediatric nephrologists participating in the Pediatric Nephrology Research Consortium meeting in April 2023. RESULTS: Sixty-four pediatric nephrologists completed the survey. Twenty-seven (42.2%) respondents reported having prescribed probiotics to pediatric kidney transplant recipients. The primary reason for probiotic use was the treatment of antibiotic-associated diarrhea (n = 20), with other reasons including recurrent Clostridium difficile infection (n = 15), general gut health promotion (n = 12), recurrent urinary tract infections (n = 8), and parental request (n = 1). Of those who prescribed probiotics, 48.1% held them during periods of neutropenia and 14.8% during central venous line use. Of the 64 respondents, 20 reported the lack of safety data as a concern for using probiotics in kidney transplant recipients. CONCLUSION: Pediatric nephrologists are increasingly prescribing probiotics to pediatric kidney transplant recipients; nevertheless, substantial practice variations exist. The paucity of safety data is a significant deterrent to probiotic use in this population.

The effects of Capn1 gene inactivation on skeletal muscle growth, development, and atrophy, and the compensatory role of other proteolytic systems.

Myofibrillar protein turnover is a key component of muscle growth and degeneration, requiring proteolytic enzymes to degrade the skeletal muscle proteins. The objective of this study was to investigate the role of the calpain proteolytic system in muscle growth development using µ-calpain knockout (KO) mice in comparison with control wild-type (WT) mice, and evaluate the subsequent effects of silencing this gene on other proteolytic systems. No differences in muscle development between genotypes were observed during the early stages of growth due to the up regulation of other proteolytic systems. The KO mice showed significantly greater m-calpain protein abundance (P < 0.01) and activity (P < 0.001), and greater caspase 3/7 activity (P < 0.05). At 30 wk of age, KO mice showed increased protein:DNA (P < 0.05) and RNA:DNA ratios (P < 0.01), greater protein content (P < 0.01) at the expense of lipid deposition (P < 0.05), and an increase in size and number of fast-twitch glycolytic muscle fibers (P < 0.05), suggesting that KO mice exhibit an increased capacity to accumulate and maintain protein in their skeletal muscle. Also, expression of proteins associated with muscle regeneration (neural cell adhesion molecule and myoD) were both reduced in the mature KO mice (P < 0.05 and P < 0.01, respectively), indicating less muscle regeneration and, therefore, less muscle damage. These findings indicate the concerted action of proteolytic systems to ensure muscle protein homeostasis in vivo. Furthermore, these data contribute to the existing evidence of the importance of the calpain system's involvement in muscle growth, development, and atrophy. Collectively, these data suggest that there are opportunities to target the calpain system to promote the growth and/or restoration of skeletal muscle mass.

Targeted gene inactivation reveals a functional role of calpain-1 in platelet spreading.

BACKGROUND: Calpains are implicated in a wide range of cellular functions including the maintenance of hemostasis via the regulation of cytoskeletal modifications in platelets. OBJECTIVES: Determine the functional role of calpain isoforms in platelet spreading. METHODS AND RESULTS: Platelets from calpain-1(-/-) mice show enhanced spreading on collagen- and fibrinogen-coated surfaces as revealed by immunofluorescence, differential interference contrast (DIC) and scanning electron microscopy. The treatment of mouse platelets with MDL, a cell permeable inhibitor of calpains 1/2, resulted in increased spreading. The PTP1B-mediated enhanced tyrosine dephosphorylation in calpain-1(-/-) platelets did not fully account for the enhanced spreading as platelets from the double knockout mice lacking calpain-1 and PTP1B showed only a partial rescue of the spreading phenotype. In non-adherent platelets, proteolysis and GTPase activity of RhoA and Rac1 were indistinguishable between the wild-type (WT) and calpain-1(-/-) platelets. In contrast, the ECM-adherent calpain-1(-/-) platelets showed higher Rac1 activity at the beginning of spreading, whereas RhoA was more active at later time points. The ECM-adherent calpain-1(-/-) platelets showed an elevated level of RhoA protein but not Rac1 and Cdc42. Proteolysis of recombinant RhoA, but not Rac1 and Cdc42, indicates that RhoA is a calpain-1 substrate in vitro. CONCLUSIONS: Potentiation of the platelet spreading phenotype in calpain-1(-/-) mice suggests a novel role of calpain-1 in hemostasis, and may explain the normal bleeding time observed in the calpain-1(-/-) mice.

Micro-calpain is essential for postmortem proteolysis of muscle proteins.

The objective of this investigation was to test the hypothesis that -calpain is largely responsible for postmortem proteolysis of muscle proteins. To accomplish this objective, we compared proteolysis of known muscle proteins in muscles of wild type and micro-calpain knockout mice during postmortem storage. Knockout mice (n = 6) were killed along with control mice (n = 6). Hind limbs were removed and stored at 4 degrees C. Muscles were dissected at 0, 1, and 3d postmortem and subsequently analyzed for degradation of nebulin, dystrophin, metavinculin, vinculin, desmin, and troponin T. In a separate experiment, hind limb muscles from knockout (n = 4) and control mice (n = 4) were analyzed at 0, 1, and 3 d postmortem using casein zymography to confirm that mu-calpain activity was knocked out in muscle and to determine whether or not m-calpain is activated in murine postmortem muscle. Cumulatively, the results of the first experiment indicated that postmortem proteolysis was largely inhibited in micro-calpain knockout mice. The results of the second experiment established the absence of micro-calpain in the muscle tissue of knockout mice and confirmed the results of an earlier study that m-calpain is active in postmortem murine muscle. The results of the current study show that even in a species in which m-calpain is activated to some extent postmortem, micro-calpain is largely responsible for postmortem proteolysis. This observation excludes a major role for any of the other members of the calpain family or any other proteolytic system in postmortem proteolysis of muscle proteins. Therefore, understanding the regulation of micro-calpain in postmortem muscle should be the focus of further research on postmortem proteolysis and tenderization of meat.

Host receptors in malaria merozoite invasion.

The clinical manifestations of Plasmodium falciparum malaria are directly linked to the blood stage of the parasite life cycle. At the blood stage, the circulating merozoites invade erythrocytes via a specific invasion pathway often identified with its dependence or independence on sialic acid residues of the host receptor. The invasion process involves multiple receptor-ligand interactions that mediate a complex series of events in a period of approximately 1 min. Although the mechanism by which merozoites invade erythrocytes is not fully understood, recent advances have put a new perspective on the importance of developing a multivalent blood stage-malaria vaccine. In this review, we highlight the role of currently identified host invasion receptors in blood-stage malaria.

Villin-type headpiece domains show a wide range of F-actin-binding affinities.

The villin-type "headpiece" domain is a modular motif found at the extreme C-terminus of larger "core" domains in over 25 cytoskeletal proteins in plants and animals. Although headpiece is classified as an F-actin-binding domain, it has been suggested that some expressed fusion-proteins containing headpiece may lack F-actin-binding in vivo. To determine the intrinsic F-actin affinity of headpiece domains, we quantified the F-actin affinity of seven headpiece domains and three N-terminal truncations, under identical in vitro conditions. The constructs are folded and adopt the native headpiece structure. However, they show a wide range of affinities that can be grouped into high, low, and nonspecific-binding categories. Computer models of the structure and charged surface potential of these headpiece domains suggest features important for high F-actin affinity. We conclude that not all headpiece domains are intrinsically F-actin-binding motifs, and suggest that the surface charge distribution may be an important element for F-actin recognition.

Reassignment of the EPB4.1 gene to 1p36 and assessment of its involvement in neuroblastomas.

OBJECTIVES: EPB4.1 has been previously mapped to human chromosome 1p33-p34.2. In contradiction to this chromosomal location, we have mapped EPB4.1-1p36 by using fluorescence in situ hybridization and radiation hybrid mapping. In neuroblastomas, deletions of the telomeric end of chromosome 1 (1p36) are the most common genetic aberration. METHODS: We investigated whether genetic aberrations of EPB4.1 can be detected in some neuroblastomas by analyzing 72 tumours for EPB4.1 mutation, expression, and alternative splicing pattern. Furthermore, EPB4.1 protein from a neuroblastoma cell line was studied for its subcellular localization. RESULTS: Sequence changes could be detected in 14 out of 72 neuroblastomas, including missense, silent, and intronic changes. Duplex RT-PCR analysis revealed a subset of 11 tumours expressing significantly low levels of EPB4.1. Significant EPB4.1 sequence changes that were detected included an exon 4 G/A missense mutation (amino acid: V/I) that was shown to be associated with absence of wild-type EPB4.1 expression (3 tumours), an exon 8 G/A missense mutation (V/M) (1 tumour), and an intronic sequence change that was shown to be associated with the presence of an aberrant transcript (1 tumour). Splicing pattern analysis revealed that all EPB4.1 transcripts from tumours exclude exon 3, a splicing pattern for generating the 135 kDa isoform. EPB4.1 cDNA cloned from a neuroblastoma cell line produced a 135-kDa protein with a cytoplasm/membrane localization. CONCLUSIONS: Out of 72 neuroblastomas we have identified 11 tumours with impaired EPB4.1 expression and 5 tumours with significant sequence changes. We also found that the 135 kDa isoform is the main EPB4.1 product in neuroblastoma. EPB4.1 cDNA from a neuroblastoma cell line produced a 135-kDa protein and displayed a cytoplasm/membrane localization in transfected cells.

Expression of chemokine receptors CXCR1 and CXCR2 during cardiopulmonary bypass.

OBJECTIVE: This study investigated the effects of cardiopulmonary bypass on neutrophil expression of chemokine receptors, CXCR1 and CXCR2, and the beta2 integrin CD11b. METHODS: Ten patients undergoing coronary artery grafting with cardiopulmonary bypass were studied. Blood samples were collected preoperatively, before bypass, at termination of bypass, and 12 to 18 hours postoperatively. In vitro studies were performed on control subjects to determine changes in the surface expression of CXCR1, CXCR2, and CD11b on stimulation with interleukin 8. Receptor expression was measured by flow cytometry. Plasma levels of interleukin 8 from the patients were determined by enzyme-linked immunoassay. RESULTS: After bypass, CXCR2 expression fell by 66% (P <.0001) and remained low postoperatively (P <.0001). CXCR1 expression persisted at preoperative levels. CD11b expression increased significantly after bypass (P <.0001), returning to prebypass levels postoperatively. In vitro studies showed a dose-related fall of both CXCR1 (P <.0001) and CXCR2 expression (P <.0001) and a significant rise in CD11b expression (P <.0001). Plasma interleukin 8 increased significantly after bypass (P <.0001), remaining elevated 12 to 18 hours postoperatively (P =.02). Correlations between interleukin 8 levels and CXCR2 expression (P <.0001) and CD11b expression (P <.03) were demonstrated. CONCLUSIONS: CXCR2 expression is significantly down-regulated after bypass; in contrast, CXCR1 expression remains unchanged. In addition, whereas interleukin 8 is an important determinant of both CXCR1 and CXCR2 expression in vitro, it only correlates with CXCR2 and CD11b expression in vivo. This has implications in the search for antagonists against CXC chemokines and their receptors.

Long-term treatment of focal segmental glomerulosclerosis in children with cyclosporine given as a single daily dose.

Cyclosporine (CsA) has been successfully used for treatment of children with focal segmental glomerulosclerosis (FSGS) and nephrotic syndrome (NS) for the last decade. Response rates of 50% to 100% have been reported using twice-daily dosing of 5 to 32 mg/kg/d, achieving trough blood levels of 70 to 500 ng/mL. Treatment has been associated with a high incidence of side effects, including nephrotoxicity, hypertension, gingival hyperplasia, and hirsutism. To determine whether once-daily low-dose CsA could minimize side effects and still induce remission, 21 children with biopsy-proven FSGS and NS, each treated with CsA, 4.6 +/- 0.8 mg/kg/d, with no predetermined target trough blood levels, were studied. Eleven of 21 children (52%) attained complete remission and 5 of 21 children (24%) attained partial remission, for a total response rate of 76%. Mean time to response was 2.8 +/- 0.8 months, and mean duration of therapy was 20.6 +/- 13.7 months. CsA dosage was tapered or stopped in 9 responders; 3 of these patients maintained remission at last follow-up 6 to 13 months later, and 6 patients relapsed at 1.5 to 18.7 months (mean, 8.7 months). Five of these 6 patients responded again when CsA therapy was restarted or the dosage was increased. Twelve of 16 responders were still being administered CsA at last follow-up 11 to 60 months (mean, 24.6 months) later. Five of 21 patients (24%) had no response to CsA during 2 to 27 months of therapy; 4 of these 5 patients developed end-stage renal disease after CsA therapy was stopped. Side effects of CsA therapy were minimal: 1 patient each developed new-onset hypertension or gingival hyperplasia, and no patient had hirsutism or nephrotoxicity. Single daily low-dose CsA appears to be effective for long-term treatment of children with FSGS and NS, with fewer side effects than twice-daily dosing.

Schwannomin isoform-1 interacts with syntenin via PDZ domains.

The neurofibromatosis type 2 gene (NF2) is involved in the pathogenesis of benign tumors of the human nervous system. The NF2 protein, called schwannomin or merlin, is inactivated in virtually all schwannomas and meningiomas. The molecular mechanisms by which schwannomin functions as a tumor suppressor is unknown but believed to involve plasma membrane-cytoskeletal interactions. Two major alternatively spliced isoforms of schwannomin differing in their C termini have been reported. Using the yeast two-hybrid system, we have identified syntenin as a binding partner for schwannomin isoform-1 (sch-1). Syntenin is an adapter protein that couples transmembrane proteoglycans to cytoskeletal components and is involved in intracellular vesicle transport. The C terminus 25 amino acids of sch-1 and the two PDZ domains of syntenin mediate their binding, and mutations introduced within the VAFFEEL region of sch-1 defined a sequence crucial for syntenin recognition. We have showed that the two proteins interacted in vitro and in vivo and localized underneath the plasma membrane. Fibroblast cells expressing heterologous antisense syntenin display alterations in the subcellular distribution of sch-1. Together, these results provide the first functional clue to the existence of schwannomin isoforms and could unravel novel pathways for the transport and subcellular localization of schwannomin in vivo.

VAM-1: a new member of the MAGUK family binds to human Veli-1 through a conserved domain.

The MAGUKs (membrane-associated guanylate kinase homologues) constitute a family of peripheral membrane proteins that function in tumor suppression and receptor clustering by forming multiprotein complexes containing distinct sets of transmembrane, cytoskeletal, and cytoplasmic signaling proteins. Here, we report the characterization of the human vam-1 gene that encodes a novel member of the p55 subfamily of MAGUKs. The complete cDNA sequence of VAM-1, tissue distribution of its mRNA, genomic structure, chromosomal localization, and Veli-1 binding properties are presented. The vam-1 gene is composed of 12 exons and spans approx. 115 kb. By fluorescence in situ hybridization the vam-1 gene was localized to 7p15-21, a chromosome region frequently disrupted in some human cancers. VAM-1 mRNA was abundant in human testis, brain, and kidney with lower levels detectable in other tissues. The primary structure of VAM-1, predicted from cDNA sequencing, consists of 540 amino acids including a single PDZ domain near the N-terminus, a central SH3 domain, and a C-terminal GUK (guanylate kinase-like) domain. Sequence alignment, heterologous transfection, GST pull-down experiments, and blot overlay assays revealed a conserved domain in VAM-1 that binds to Veli-1, the human homologue of the LIN-7 adaptor protein in Caenorhabditis. LIN-7 is known to play an essential role in the basolateral localization of the LET-23 tyrosine kinase receptor, by linking the receptor to LIN-2 and LIN-10 proteins. Our results therefore suggest that VAM-1 may function by promoting the assembly of a Veli-1 containing protein complex in neuronal as well as epithelial cells.

Disruption of the mouse mu-calpain gene reveals an essential role in platelet function.

Conventional calpains are ubiquitous calcium-regulated cysteine proteases that have been implicated in cytoskeletal organization, cell proliferation, apoptosis, cell motility, and hemostasis. There are two forms of conventional calpains: the mu-calpain, or calpain I, which requires micromolar calcium for half-maximal activation, and the m-calpain, or calpain II, which functions at millimolar calcium concentrations. We evaluated the functional role of the 80-kDa catalytic subunit of mu-calpain by genetic inactivation using homologous recombination in embryonic stem cells. The mu-calpain-deficient mice are viable and fertile. The complete deficiency of mu-calpain causes significant reduction in platelet aggregation and clot retraction but surprisingly the mutant mice display normal bleeding times. No detectable differences were observed in the cleavage pattern and kinetics of calpain substrates such as the beta3 subunit of alphaIIbbeta3 integrin, talin, and ABP-280 (filamin). However, mu-calpain null platelets exhibit impaired tyrosine phosphorylation of several proteins including the beta3 subunit of alphaIIbbeta3 integrin, correlating with the agonist-induced reduction in platelet aggregation. These results provide the first direct evidence that mu-calpain is essential for normal platelet function, not by affecting the cleavage of cytoskeletal proteins but by potentially regulating the state of tyrosine phosphorylation of the platelet proteins.

A cysteine protease activity from Plasmodium falciparum cleaves human erythrocyte ankyrin.

The malaria parasite Plasmodium falciparum undergoes distinct morphologic changes during its 48-h life cycle inside human red blood cells. Parasite proteinases appear to play important roles at all stages of the erythrocytic cycle of human malaria. Proteases involved in erythrocyte rupture and invasion are possibly required to breakdown erythrocyte membrane skeleton. To identify such proteases, soluble cytosolic extract of isolated trophozoites/schizonts was incubated with erythrocyte membrane ghosts or spectrin-actin depleted inside-out vesicles, which were then analyzed by SDS-PAGE. In both cases, a new protein band of 155 kDa was detected. The N-terminal peptide sequencing established that the 155 kDa band represents truncated ankyrin. Immunoblot analysis using defined monoclonal antibodies confirmed that ankyrin was cleaved at the C-terminus. While the enzyme preferentially cleaved ankyrin, degradation of protein 4.1 was also observed at high concentrations of the enzyme. The optimal activity of the purified enzyme, using ankyrin as substrate, was observed at pH 7.0-7.5, and the activity was strongly inhibited by standard inhibitors of cysteine proteinases (cystatin, NEM, leupeptin, E-64 and MDL 28 170), but not by inhibitors of aspartic (pepstatin) or serine (PMSF, DFP) proteinases. Furthermore, we demonstrate that protease digestion of ankyrin substantially reduces its interaction with ankyrin-depleted membrane vesicles. Ektacytometric measurements showed a dramatic increase in the rate of fragmentation of ghosts after treatment with the protease. Although the role of ankyrin cleavage in vivo remains to be determined, based on our findings we postulate that the parasite-derived cysteine protease activity cleaves host ankyrin thus weakening the ankyrin-band 3 binding interactions and destabilizing the erythrocyte membrane skeleton, which, in turn, facilitates parasite release. Further characterization of the enzyme may lead to the development of novel antimalarial drugs.

hCASK and hDlg associate in epithelia, and their src homology 3 and guanylate kinase domains participate in both intramolecular and intermolecular interactions.

Membrane-associated guanylate kinase (MAGUK) proteins act as molecular scaffolds organizing multiprotein complexes at specialized regions of the plasma membrane. All MAGUKs contain a Src homology 3 (SH3) domain and a region homologous to yeast guanylate kinase (GUK). We showed previously that one MAGUK protein, human CASK (hCASK), is widely expressed and associated with epithelial basolateral plasma membranes. We now report that hCASK binds another MAGUK, human discs large (hDlg). Immunofluorescence microscopy demonstrates that hCASK and hDlg colocalize at basolateral membranes of epithelial cells in small and large intestine. These proteins co-precipitate from lysates of an intestinal cell line, Caco-2. The GUK domain of hCASK binds the SH3 domain of hDlg in both yeast two-hybrid and fusion protein binding assays, and it is required for interaction with hDlg in transfected HEK293 cells. In addition, the SH3 and GUK domains of each protein participate in intramolecular binding that in vitro predominates over intermolecular binding. The SH3 and GUK domains of human p55 display the same interactions in yeast two-hybrid assays as those of hCASK. Not all SH3-GUK interactions among these MAGUKs are permissible, however, implying specificity to SH3-GUK interactions in vivo. These results suggest MAGUK scaffold assembly may be regulated through effects on intramolecular SH3-GUK binding.

Plasmodium falciparum erythrocyte membrane protein 1 is anchored to the actin-spectrin junction and knob-associated histidine-rich protein in the erythrocyte skeleton.

A distinctive pathological feature of Plasmodium falciparum malaria is the endothelial attachment of erythrocytes infected with mature asexual-stage parasites in microvessels of the major organs. Electron-dense protrusions described as knobs are displayed on the surface of parasitized erythrocytes and act as attachment points in cytoadherence. Parasite-encoded knob-associated histidine-rich protein (KAHRP) is a major component of knobs found on the cytoplasmic side of the host cell membrane. P. falciparum erythrocyte membrane protein 1 (PfEMP1) is a family of parasite-encoded cytoadherence receptors localized to knobs on the surface of parasitized erythrocytes. Despite its high antigenic diversity, PfEMP1 has a remarkably conserved cytoplasmic domain. We demonstrate in this study that the cytoplasmic domain of PfEMP1 (VAR(CD)) binds to host spectrin and actin and to full-length KAHRP in vitro. Apparent dissociation constants determined for VAR(CD)/F-actin and VAR(CD)/KAHRP interactions are 44.9+/-6.4 and 10. 7+/-2.2 nM, respectively. Further, we provide evidence that KAHRP polypeptides self-associate in solution to form structures similar to knobs and show binding of self-associated KAHRP clusters to spectrin-actin-protein 4.1 complexes. Findings in this study suggest that PfEMP1 is localized to the knob in P. falciparum-infected erythrocytes by binding to the host spectrin-actin junction and to self-associated KAHRP through its conserved cytoplasmic domain.

GAKIN, a novel kinesin-like protein associates with the human homologue of the Drosophila discs large tumor suppressor in T lymphocytes.

Reorganization of the cortical cytoskeleton is a hallmark of T lymphocyte activation. Upon binding to antigen presenting cells, the T cells rapidly undergo cytoskeletal re-organization thus forming a cap at the cell-cell contact site leading to receptor clustering, protein segregation, and cellular polarization. Previously, we reported cloning of the human lymphocyte homologue of the Drosophila Discs Large tumor suppressor protein (hDlg). Here we show that a novel protein termed GAKIN binds to the guanylate kinase-like domain of hDlg. Affinity protein purification, peptide sequencing, and cloning of GAKIN cDNA from Jurkat J77 lymphocytes identified GAKIN as a novel member of the kinesin superfamily of motor proteins. GAKIN mRNA is ubiquitously expressed, and the predicted amino acid sequence shares significant sequence similarity with the Drosophila kinesin-73 motor protein. GAKIN sequence contains a motor domain at the NH(2) terminus, a central stalk domain, and a putative microtubule-interacting sequence called the CAP-Gly domain at the COOH terminus. Among the MAGUK superfamily of proteins examined, GAKIN binds to the guanylate kinase-like domain of PSD-95 but not of p55. The hDlg and GAKIN are localized mainly in the cytoplasm of resting T lymphocytes, however, upon CD2 receptor cross-linking the hDlg can translocate to the lymphocyte cap. We propose that the GAKIN-hDlg interaction lays the foundation for a general paradigm of coupling MAGUKs to the microtubule-based cytoskeleton, and that this interaction may be functionally important for the intracellular trafficking of MAGUKs and associated protein complexes in vivo.

Sleep-related breathing disorders following discharge from intensive care.

OBJECTIVES: To determine the incidence of sleep-related breathing disorders and nocturnal hypoxaemia in patients discharged from ICU following prolonged mechanical ventilation. DESIGN: Prospective, consecutive patient observational study. SETTING: The medical and surgical wards of a University Hospital. PATIENTS AND PARTICIPANTS: Fifteen consecutive, adult patients discharged from the ICU who had received more than 48 h of mechanical ventilation were studied. Ten healthy volunteers acted as controls. MEASUREMENTS AND RESULTS: Overnight, multi-channel pneumographic studies were performed on all patients and controls. Chest and abdominal wall movement, air flow, oxygen saturation and snoring were continuously recorded. Data was analysed by both visual inspection of the traces and by computer-based algorithms. An apnoea/hypopnoea index was calculated for each patient and volunteer. Volunteers had an apnoea/hypopnoea index of less than 5 and had no episodes of nocturnal oxygen desaturation (SaO2 < 90%). Despite oxygen therapy 13/15 patients had episodes of desaturation and 9/15 spent more than 2 h with an SaO2 < 90%. Eleven patients had an abnormal apnoea/hypopnoea index (range 5-34 events/h). Four patients had predominantly obstructive events while 7 primarily had hypopnoeas. CONCLUSIONS: Significant overnight oxygen desaturation is common in patients discharged from ICU who have received prolonged mechanical ventilation. This group also has a significant incidence of sleep-related breathing disorders and this mechanism is likely to be important in the pathogenesis of the hypoxaemia.