Associations between glucagon prescribing, hospital admissions for hypoglycaemia and continuous glucose monitoring metrics in adults with type 1 diabetes.

AIMS: To assess features associated with glucagon prescribing and hospital admissions with hypoglycaemia in type one diabetes. METHODS: Observational study of 4462 adults. Outcome measures were features associated with glucagon prescriptions and predictors of hospital admissions with hypoglycaemia and high levels of glucagon prescribing. RESULTS: 74 % did not collect any glucagon prescriptions and 2.7 % collected >6 over 3.5 years. Hospital admission with hypoglycaemia (P = 0.032), impaired awareness (P = 0.049) and female sex (P < 0.001) were associated with glucagon collection. More frequent prescribing of glucagon was associated with diabetes duration (P < 0.001) and socioeconomic deprivation (P < 0.001). Higher average glucose (P = 0.047), higher time above 13.9 mM (P = 0.008) and higher SD (P = 0.002) were associated with glucagon prescribing. Diabetes duration (P < 0.001) and HbA1c (P < 0.001) were higher in people with hospitalised hypoglycaemia. Higher time above 13.9 mM (P = 0.004) and SD glucose (P < 0.001) were most clearly associated with hospitalised hypoglycaemia. CONCLUSIONS: A minority of people with type 1 diabetes have access to glucagon suggesting more could be done to better target this treatment. Individuals with risk factors and those with frequent glucagon prescriptions should be identified for interventions known to reduce hypoglycaemia.

Changes in continuous glucose monitoring metrics and predictors of improvement 12 months after conversion from Freestyle Libre to Freestyle Libre 2.

AIMS: We sought to assess whether conversion from Freestyle Libre to Freestyle Libre 2 (with low and high glucose alert functions) was associated with improved glucose metrics. RESEARCH DESIGN AND METHODS: A prospective observational study to assess changes in CGM metrics in 672 adults with type 1 diabetes when converting to Freestyle Libre 2. Secondary outcomes included predictors of reduction in time below range (TBR) and increase in time in range (TIR). RESULTS: TBR fell by a median of 1.0% (IQR -2.7 to 0.3, p < 0.001) after 12 months and TIR decreased by 1.0% (-8.7 to 6.0, p = 0.004). TIR did not fall in people using high glucose alerts (p = 0.353). Average duration of low glucose events (<3.9 mmoL/L) fell by 10 min (-46 to 18, p < 0.001). Significant improvements in TIR (p = 0.029) and time above 13.9 mM (p = 0.002) were observed in those using high glucose alerts. Alert threshold settings were not associated with glycaemic response; however, low alert use was independently associated with a fall in TBR of ≥0.5% (HR 1.9 [95% CI 1.2-3.1], p = 0.009) and high alert use was independently associated with a rise in TIR of ≥5% (HR 1.6 [95% CI 1.0-2.5], p = 0.043) at 12 months. CONCLUSIONS: Conversion to Freestyle Libre 2 was associated with significant improvements in low glucose metrics. Alert function users were more likely to see improvements across all CGM metrics. Challenges remain in encouraging alert use, helping users set optimal alert thresholds and optimizing response to alerts.

Regularized robust estimation in binary regression models.

In this paper, we investigate robust parameter estimation and variable selection for binary regression models with grouped data. We investigate estimation procedures based on the minimum-distance approach. In particular, we employ minimum Hellinger and minimum symmetric chi-squared distances criteria and propose regularized minimum-distance estimators. These estimators appear to possess a certain degree of automatic robustness against model misspecification and/or for potential outliers. We show that the proposed non-penalized and penalized minimum-distance estimators are efficient under the model and simultaneously have excellent robustness properties. We study their asymptotic properties such as consistency, asymptotic normality and oracle properties. Using Monte Carlo studies, we examine the small-sample and robustness properties of the proposed estimators and compare them with traditional likelihood estimators. We also study two real-data applications to illustrate our methods. The numerical studies indicate the satisfactory finite-sample performance of our procedures.

Substantial HbA1c Reduction Following Intermittent-Scanning Continuous Glucose Monitoring Was Not Associated With Early Worsening of Retinopathy in Type 1 Diabetes.

BACKGROUND: Early worsening of diabetic retinopathy (EWDR) was observed in the intensively treated arm of the Diabetes Control and Complications Trial (DCCT) before long-term benefits accrued. We sought to assess whether there may be an increased risk of EWDR in high-risk individuals following intermittent-scanning continuous glucose monitoring (iscCGM) commencement. METHODS: An observational study of 139 individuals with type 1 diabetes ≥5 years duration and with baseline HbA1c >75 mmol/mol (9.0%). This cohort was stratified by subsequent HbA1c response to iscCGM (best responders and non-responders). Pan-retinal photocoagulation (PRP), worsening retinopathy status and new development of retinopathy were compared between groups. RESULTS: HbA1c change was -23 mmol/mol (IQR -32 to -19) (-2.1% [-2.9 to -1.8]) in responders and +6 mmol/mol (2-12) (+0.6 [0.2-1.1]) in non-responders (P < .001). There was no difference in subsequent PRP between responders (14.1%) and non-responders (10.3%, P = .340). Baseline HbA1c (HR 1.052 per mmol/mol, P = .002) but not response category (HR 1.244, P = .664) was independently associated with the risk of requiring PRP. Worsening of retinopathy was not different between responders (16.9%) and non-responders (20.6%, P = .577), and the same was true with respect to new development of retinopathy (33.3% vs 31.8%, P = .919). CONCLUSIONS: In a cohort enriched for risk of diabetic retinopathy, reduction in HbA1c did not result in an increased risk of PRP, worsening retinopathy, or new development of retinopathy. These findings offer reassurance that substantial reduction in HbA1c is not independently associated with early worsening of diabetic eye disease in iscCGM users.

Socioeconomic deprivation, technology use, C-peptide, smoking and other predictors of glycaemic control in adults with type 1 diabetes.

AIMS: Intensive glycaemic control is associated with substantial health benefits in people with type 1 diabetes. We sought to examine clinical and demographic factors associated with meeting glycaemic targets in type 1 diabetes. METHODS: We conducted a cross-sectional analysis of 4594 individuals with type 1 diabetes. The primary outcome of the study was assessing factors associated with meeting HbA1c targets. Secondary endpoints included factors associated with continuous subcutaneous insulin infusion (CSII) use and persistent C-peptide secretion. RESULTS: Socioeconomic deprivation was strongly associated with a lower likelihood of achieving an HbA1c <58 mmol/mol (7.5%) (20% in the most deprived quintile vs. 40% in the least deprived, p < 0.001). In multivariate analysis, absence of smoking history (OR 3.06, p < 0.001), flash monitoring (OR 1.49, p < 0.001), CSII (1.43, p = 0.022) and longer diabetes duration (OR 1.02 per year, p = 0.004) were independently associated with achieving HbA1c <58 mmol/mol (7.5%), whereas increasing age (OR 0.99 per year, p = 0.004) and C-peptide <50 pM (OR 0.58, p < 0.001) were associated with a lower likelihood of meeting this target. Low C-peptide (<50 pM) was less likely in men (OR 0.55, p < 0.001) and never smokers (0.44, p < 0.001) in multivariate analysis. CONCLUSIONS: Lower levels of deprivation, non-smoking, higher C-peptide, technology use, lower BMI and male gender were all associated with a higher likelihood of meeting HbA1c targets. Access to proven diabetes treatments is lower in the most deprived individuals. Urgent efforts are required to provide treatments which are effective across the socioeconomic gradient.

HbA1c response and hospital admissions following commencement of flash glucose monitoring in adults with type 1 diabetes.

INTRODUCTION: Our aim was to assess the effect of introducing flash monitoring in adults with type 1 diabetes with respect to change in hemoglobin A1c (HbA1c) and frequency of hospital admissions. RESEARCH DESIGN AND METHODS: Prospective observational study of adults with type 1 diabetes in our center, in whom a prescription for a flash monitoring sensor was collected. Primary outcome was change in HbA1c between 2016 and after flash monitoring. Rates of hospital admission were compared between the first year after flash monitoring and the corresponding 12-month period 2 years earlier. RESULTS: Approximately half of all adults with type 1 diabetes, attending our center, collected prescriptions for flash monitoring sensors (n=2216). Median fall in HbA1c was -1 (-0.1) mmol/mol (%) (p<0.001) and was greatest in those with baseline HbA1c >75 (9.0) mmol/mol (%): -10 (-0.9) mmol/mol (%), p<0.001. 43% of those with a baseline HbA1c >53 mmol/mol (7%) experienced a ≥5 mmol/mol (0.5%) fall in HbA1c. In addition to higher HbA1c, early commencement within 1 month of NHS-funded flash monitoring (p<0.001), and male gender (p=0.013) were associated with a fall in HbA1c of ≥5 (0.5) mmol/mol (%). Socioeconomic deprivation (p=0.009) and collecting fewer than 2 sensors per month (p=0.002) were associated with lack of response. Overall, hospital admissions did not change but an increase in admissions for hypoglycemia was observed (1.1% vs 0.3%, p=0.026). CONCLUSIONS: Flash monitoring is associated with reduction in HbA1c in individuals with HbA1c >58 mmol/mol. Numerous clinical features are independently associated with HbA1c response. An increase in hypoglycemia admissions occurred following flash monitoring.

Efficient robust doubly adaptive regularized regression with applications.

We consider the problem of estimation and variable selection for general linear regression models. Regularized regression procedures have been widely used for variable selection, but most existing methods perform poorly in the presence of outliers. We construct a new penalized procedure that simultaneously attains full efficiency and maximum robustness. Furthermore, the proposed procedure satisfies the oracle properties. The new procedure is designed to achieve sparse and robust solutions by imposing adaptive weights on both the decision loss and the penalty function. The proposed method of estimation and variable selection attains full efficiency when the model is correct and, at the same time, achieves maximum robustness when outliers are present. We examine the robustness properties using the finite-sample breakdown point and an influence function. We show that the proposed estimator attains the maximum breakdown point. Furthermore, there is no loss in efficiency when there are no outliers or the error distribution is normal. For practical implementation of the proposed method, we present a computational algorithm. We examine the finite-sample and robustness properties using Monte Carlo studies. Two datasets are also analyzed.

'Training gap' - the present state of higher specialty training in adolescent and young adult health in medical specialties in the UK.

Young people often experience worse health outcomes and more dissatisfaction with healthcare compared with other age groups. This survey sought to determine the state of adolescent and young adult health training across medical specialties in the UK. An online questionnaire was distributed to higher specialty trainees in adult medical specialties. Training in adolescent/young adult health/transition was rated as minimal/non-existent by 70/73% of respondents, respectively; 52% reported that they had received no formal training and 61% had never attended a dedicated young person's or transition clinic. The most significant barrier to delivering good adolescent and young adult healthcare was felt to be lack of training to deal with adolescent issues. This survey has identified a 'training gap'; a lack of preparation to meet the specific care needs of the adolescent and young adult population. Improved interventions are required to help drive improvement in care for young people in the UK.

Frequent nasopharyngeal suctioning as a risk factor associated with neonatal coagulase-negative staphylococcal colonisation and sepsis.

INTRODUCTION: This case-control study aimed to determine whether catheter use was significantly associated with coagulase-negative staphylococci (CoNS) colonisation and/or sepsis in neonates. METHODS: Weekly swabs of the nose, umbilicus, rectum, wounds, eye discharge and intravenous catheter tips (after removal) of infants admitted to the neonatal intensive care unit of Universiti Kebangsaan Malaysia Medical Centre, Malaysia, were cultured. CoNS sepsis was diagnosed if pure growth of CoNS was cultured from the peripheral blood specimen of symptomatic infants. For each infant with CoNS colonisation or sepsis, a control infant was retrospectively and randomly selected from unaffected infants in the ward. Multivariate analyses were performed to determine whether catheter use was a significant risk factor. RESULTS: CoNS colonisation was detected in 113 (8.7%) infants. CoNS sepsis was found in 12 (10.6%) infants with CoNS colonisation and 7 (0.6%) infants without CoNS colonisation. Multivariate analysis showed that the following were significantly associated with CoNS colonisation: conjunctivitis (adjusted odds ratio [OR] 8.2, 95% confidence interval [CI] 1.9­34.8, p = 0.005); central venous catheters (adjusted OR 5.8, 95% CI 1.9­17.8, p = 0.002); and nasopharyngeal and/or oral suctioning more than twice in the 48 hours before positive culture (adjusted OR 7.3, 95% CI 3.3­16.2, p < 0.001). Exposure to frequent nasopharyngeal and/or oral suctioning (adjusted OR 20.8, 95% CI 3.5­125.3, p = 0.001) was the only significant factor associated with CoNS sepsis. CONCLUSION: Infants requiring more than two nasopharyngeal and/or oral suctions in the previous 48 hours were found to have a higher risk of developing CoNS colonisation and sepsis.

Prevalence of impaired awareness of hypoglycemia and identification of predictive symptoms in children and adolescents with type 1 diabetes.

BACKGROUND: In children with type 1 diabetes mellitus (T1DM) the prevalence of impaired awareness of hypoglycemia (IAH) is uncertain. This study aimed to ascertain this with greater precision. Secondary aims were to assess symptoms of hypoglycemia and which of these best predict awareness of hypoglycemia in children. METHODS: Questionnaires were completed by 98 children with T1DM (mean age 10.6 yr) and their parent(s); hospital admission data for the previous year were collected. Awareness of hypoglycemia was assessed using two questionnaire-based methods that have been validated in adults. For 4 wk, participants performed routine blood glucose measurements and completed questionnaires after each episode of hypoglycemia. Principal components analysis determined how symptoms correlate; multinomial logistic regression models identified which symptom aggregate best predicted awareness status. RESULTS: The 'Gold' questionnaire classified a greater proportion of the participants as having IAH than the 'Clarke' questionnaire (68.4 vs. 22.4%). Using the 'Clarke' method, but not the 'Gold' method, children with IAH were younger and more likely to require external assistance or hospital admission. Most aged ≥9 yr (98.6%) were able to self-assess awareness status accurately. Puberty and increasing age, augmented symptom scores; duration of diabetes and glycemic control had no effect. In contrast to adults, behavioral symptoms were the best predictors of awareness status. CONCLUSIONS: IAH affects a substantial minority of children and impending hypoglycemia may be heralded by behavioral symptoms. The 'Clarke' method was more effective at identifying those at increased risk and could be used as a screening tool.

Brain herniation in a neonate.

Brain herniation is generally thought to be unlikely to occur in newborns due to the presence of the patent fontanelles and cranial sutures. A review of the literature published from 1993 to 2008 via MEDLINE search revealed no reports on neonatal brain herniation from intracranial tumour. We report a preterm Malay male infant born via elective Caesarean section for antenatally diagnosed intracerebral tumour, which subsequently developed herniation. Cerebral magnetic resonance imaging showed features that were compatible with a large complex intracranial tumour causing mass effect and gross hydrocephalus. Tumour excision was scheduled when the infant was two weeks old. Unfortunately, on the morning of the surgery, he developed signs of brain herniation and had profuse tumour haemorrhage during the attempted excision. Histopathological examination revealed an embryonal tumour, possibly an atypical rhabdoid/teratoid tumour. This case illustrates that intracranial tumours in newborns can herniate and should therefore be closely monitored.

Modeling the effects of hypoglycemia on a two-choice task in adult humans.

OBJECTIVE: Previous research has demonstrated that hypoglycemia causes reaction times to be slower and more variable. Reaction time tests, however, use multiple cognitive and noncognitive processes. This study is the first to use a validated sequential sampling model (diffusion model) applied to results obtained from a simple 2-choice task in adult humans to assess the effects of hypoglycemia on the basic parameters of decision making. METHOD: Fourteen adult volunteers were tested on a numerosity discrimination task with and without reduced blood glucose concentrations. The results were analyzed with a model that dissects the components of processing that underlie decisions: the quality of the information on which a decision is based (drift rate), the critical amount of evidence that must be accumulated before a decision is made (boundary separation), and the time taken by nondecision processes. RESULTS: Hypoglycemia resulted in a reduction of mean drift rate from 0.290 to 0.211, t(13) = 4.10, p < .05. No effect of experimental state was observed on the amount of evidence required to make a decision or peripheral and motor processes. CONCLUSION: This study locates the precise processing deficit associated with hypoglycemia and provides further understanding of the precise cognitive effect of hypoglycemia. Further research into the amelioration of these effects is required.

Effects of acute insulin-induced hypoglycemia on indices of inflammation: putative mechanism for aggravating vascular disease in diabetes.

OBJECTIVE: To examine the effects of acute insulin-induced hypoglycemia on inflammation, endothelial dysfunction, and platelet activation in adults with and without type 1 diabetes. RESEARCH DESIGN AND METHODS: We studied 16 nondiabetic adults and 16 subjects with type 1 diabetes during euglycemia (blood glucose 4.5 mmol/l) and hypoglycemia (blood glucose 2.5 mmol/l). Markers of inflammation, thrombosis, and endothelial dysfunction (soluble P-selectin, interleukin-6, von Willebrand factor [vWF], tissue plasminogen activator [tPA], high-sensitivity C-reactive protein [hsCRP], and soluble CD40 ligand [sCD40L]) were measured; platelet-monocyte aggregation and CD40 expression on monocytes were determined using flow cytometry. RESULTS: In nondiabetic participants, platelet activation occurred after hypoglycemia, with increments in platelet-monocyte aggregation and P-selectin (P

Inhaled nitric oxide and intravenous magnesium sulphate for the treatment of persistent pulmonary hypertension of the newborn.

INTRODUCTION: The aim of this study was to compare the response and survival rates of term infants with persistent pulmonary hypertension of the newborn (PPHN) on high frequency oscillatory ventilation (HFOV) treated with either inhaled nitric oxide (iNO) or intravenous magnesium sulphate (MgSO4). METHODS: This was a randomised controlled study. The inclusion criteria were infants with respiratory distress, oxygen index equal to or greater than 25 despite HFOV support, and echocardiographic evidence of PPHN. Infants in the MgSO4 group (n is 13) were loaded with MgSO4 200 mg/kg infused over half an hour, followed by continuous infusion at 50-150 mg/kg/hour to attain a serum magnesium level of 5.0-7.0 mmol/L. Infants in the iNO group (n is 12) were administered nitric oxide at an initial concentration of 20 ppm. Analysis was done on an intention-to-treat basis. RESULTS: There was no significant difference in the median age when the vasodilators were commenced (MgSO4 group: 14.0 hours, interquartile range [IQR]: 7.5, 27.0; iNO group: 14.8 hours, IQR: 12.5, 35.3, p is 0.8). There was no significant difference in the proportion of infants who responded primarily to either vasodilator (MgSO4: 23.3%, iNO: 33.3%, p is 1.0) . After switching over to iNO following a failed MgSO4 therapy, a significantly higher proportion (9 out of 10) of the non-respondents in the MgSO4 group recovered from PPHN and survived compared to the non-respondents in the iNO group (1 out of 8) who switched over to intravenous MgSO4 (p is less than 0.03). CONCLUSION: Infants who were administered iNO following a failed MgSO4 therapy were associated with a better outcome than those who were administered MgSO4 following a failed iNO therapy.

The usefulness of early ultrasonography, electroencephalography and clinical parameters in predicting adverse outcomes in asphyxiated term infants.

INTRODUCTION: The early identification of asphyxiated infants at high risk of adverse outcomes and the early selection of those who might benefit from neuroprotective therapies are required. A prospective observational study was conducted to determine if there were any early clinical, neuroimaging or neurophysiological parameters that might predict the outcome in term newborns with asphyxia. METHODS: 44 term newborns with acute asphyxia had a cranial ultrasonography (US), electroencephalography (EEG) and clinical examination performed between three and eight hours of life to determine the parameters that might predict outcome. US findings were classified as normal or abnormal (ventricular dilatation or compression and/or focal/diffuse echogenicities). EEG background activity was classified into two categories: normal/mildly abnormal/intermediate, or severely abnormal (low voltage activity or "suppression-burst"). An intrapartum score (based on graded abnormalities of foetal heart monitoring, umbilical arterial base deficit and five-minute Apgar score) and a hypoxic ischaemic encephalopathy (HIE) score (based on graded abnormalities of the neurological and respiratory status at 3-8 hours of life) was also obtained. RESULTS: At one year of life, eight infants had died, six had defaulted follow-up, five had major impairment, two minor impairment and 23 were normal. On univariate analysis, poor outcome (death or major impairment) was associated with abnormal cranial US, severely abnormal EEG and a high HIE score (greater than or equal to 15). The positive predictive value was 54.5, 100 and 100 percent, respectively, while the negative predictive value was 93.8, 80.6 and 80.6 percent, respectively. Combining these factors did not improve the predictive values. CONCLUSION: There was no added advantage in combining EEG or US parameters over a clinical neurological scoring system alone in predicting the outcome of asphyxiated term newborns.

Effects of acute insulin-induced hypoglycemia on spatial abilities in adults with type 1 diabetes.

OBJECTIVE To examine the effects of acute insulin-induced hypoglycemia on spatial cognitive abilities in adult humans with type 1 diabetes. RESEARCH DESIGN AND METHODS Sixteen adults with type 1 diabetes underwent two counterbalanced experimental sessions: euglycemia (blood glucose 4.5 mmol/l [81 mg/dl]) and hypoglycemia (2.5 mmol/l [45 mg/dl]). Arterialized blood glucose levels were maintained using a hyperinsulinemic glucose clamp technique. During each session, subjects underwent detailed assessment of spatial abilities from the Kit of Factor-Referenced Cognitive Tests and two tests of general cognitive function. RESULTS Spatial ability performance deteriorated significantly during hypoglycemia. Results for the Hidden Patterns, Card Rotations, Paper Folding, and Maze Tracing tests were all impaired significantly (P < or = 0.001) during hypoglycemia, as were results for the Cube Comparisons Test (P = 0.03). The Map Memory Test was not significantly affected by hypoglycemia. CONCLUSIONS Hypoglycemia is a common side effect of insulin therapy in individuals with type 1 diabetes, and spatial abilities are of critical importance in day-to-day functioning. The deterioration in spatial abilities observed during modest experimental hypoglycemia provides novel information on the cerebral hazards of hypoglycemia that has potential relevance to everyday activities.

A quality assurance study on the administration of medication by nurses in a neonatal intensive care unit.

INTRODUCTION: This study aimed to determine the rates of non-adherence to standard steps of medication administration and medication administration errors committed by registered nurses in a neonatal intensive care unit before and after intervention. METHODS: A baseline assessment of compliance with ten standard medication administration steps by neonatal intensive care unit nurses was carried out over a two-week period. Following this, a re-education programme was launched. Three months later, they were re-assessed similarly. RESULTS: The baseline assessment showed that the nurses did not carry out at least one of the ten standard administrative steps during the administration of 188 medication doses. The most common steps omitted were having another nurse to witness drug administration (95 percent); labelling of individual medication prepared prior to administration (88 percent), checking prescription charts against patients' identification prior to administration (85 percent) and visually inspecting a patient's identification tag (71 percent) . Medication administration errors occurred in 31 percent (59/188) of doses administered, all due to imprecise timing of medication administration. There were no resultant adverse outcomes. Following implementation of remedial measures, there was a significant reduction in non-adherence of seven of the ten medication administration steps and the rate of medication administration errors (p-value is less than 0.001). However, in 94 percent of doses administered, the nurses still did not get a witness to countercheck calculations of drug dosages before administration. CONCLUSION: Non-compliance with the standard practice of medication administration by nurses is common but can be improved by continuing re-education and monitoring, plus the implementation of a standard operating procedure.

Vascular disease and diabetes: is hypoglycaemia an aggravating factor?

Acute hypoglycaemia provokes profound physiological changes affecting the cardiovascular system and several haematological parameters, principally as a consequence of sympatho-adrenal activation and counter-regulatory hormonal secretion. Many of these responses have an important role in protecting the brain from neuroglycopenia, through altering regional blood flow and promoting metabolic changes that will restore blood glucose to normal. In healthy young adults the cardiovascular effects are transient and have no obvious detrimental consequences. However, some of the effected changes are potentially pathophysiological and in people with diabetes who have developed endothelial dysfunction, they may have an adverse impact on a vasculature that is already damaged. The acute haemodynamic and haematological changes may increase the risk of localized tissue ischaemia, and major vascular events can certainly be precipitated by acute hypoglycaemia. These include myocardial and cerebral ischaemia and occasionally infarction. Established diabetic retinopathy often deteriorates after strict glycaemic control is instituted, the latter being associated with a threefold increase in frequency of severe hypoglycaemia, and enhanced exposure to mild hypoglycaemia. The possible mechanisms underlying these hypoglycaemia-induced effects include haemorrheological changes, white cell activation, vasoconstriction, and the release of inflammatory mediators and cytokines. The concept that acute hypoglycaemia could aggravate vascular complications associated with diabetes is discussed in relation to evolving comprehension of the pathogenesis of atherosclerosis and blood vessel disease.