An interpretable machine learning model for the prevention of contrast-induced nephropathy in patients undergoing lower extremity endovascular interventions for peripheral arterial disease.

BACKGROUND: Contrast-induced nephropathy (CIN) is a postprocedural complication associated with increased morbidity and mortality. An important risk factor for development of CIN is renal impairment. Identification of patients at risk for acute renal failure will allow physicians to make appropriate decisions to minimize the incidence of CIN. We developed a machine learning model to stratify risk of acute renal failure that may assist in mitigating risk for CIN in patients with peripheral artery disease (PAD) undergoing endovascular interventions. METHODS: We utilized the American College of Surgeons National Surgical Quality Improvement Program database to extract clinical and laboratory information associated with 14,444 patients who underwent lower extremity endovascular procedures between 2011 and 2018. Using 11,604 cases from 2011 to 2017 for training and 2840 cases from 2018 for testing, we developed a random forest model to predict risk of 30-day acute renal failure following infra-inguinal endovascular procedures. RESULTS: Eight variables were identified as contributing optimally to model predictions, the most important being diabetes, preoperative BUN, and claudication. Using these variables, the model achieved an area under the receiver-operating characteristic (AU-ROC) curve of 0.81, accuracy of 0.83, sensitivity of 0.67, and specificity of 0.74. The model performed equally well on white and nonwhite patients (Delong p-value = 0.955) and patients age < 65 and patients age ≥ 65 (Delong p-value = 0.659). CONCLUSIONS: We develop a model that fairly and accurately stratifies 30-day acute renal failure risk in patients undergoing lower extremity endovascular procedures for PAD. This model may assist in identifying patients who may benefit from strategies to prevent CIN.

Long-term efficacy and safety of asenapine or olanzapine in patients with schizophrenia or schizoaffective disorder: an extension study.

INTRODUCTION: Safety and efficacy results, collected in schizophrenia and schizoaffective disorder patients treated for up to nearly 3 years, are presented for asenapine and olanzapine. RESULTS: Patients completing a 52-week randomized double-blind core study on flexible-dose asenapine (5 or 10 mg BID) or olanzapine (10 or 20 mg QD) could continue treatment until study blind was broken.290 patients on asenapine and 150 on olanzapine continued treatment for variable lengths of time [mean ± SD (range) 311.0 ± 146.1 (10 - 653) d and 327.4 ± 139.6 (15 - 631) d, respectively]. Adverse event (AE) incidence was lower during the extension (asenapine, 62%; olanzapine, 55%) than during the core study (78%, 80%). In both groups, body weight increase and incidence of extrapyramidal AEs were negligible during the extension. Mean PANSS total score changes during first year of treatment were - 37.0 for asenapine and - 35.3 for olanzapine, with further changes of 1.6 for asenapine and - 0.8 for olanzapine at the extension study endpoint. CONCLUSIONS: Clinical stability on asenapine as well as olanzapine was maintained, with few recurrent or newly emerging AEs beyond 1 year of treatment.

Risk stratification with explainable machine learning for 30-day procedure-related mortality and 30-day unplanned readmission in patients with peripheral arterial disease.

Predicting 30-day procedure-related mortality risk and 30-day unplanned readmission in patients undergoing lower extremity endovascular interventions for peripheral artery disease (PAD) may assist in improving patient outcomes. Risk prediction of 30-day mortality can help clinicians identify treatment plans to reduce the risk of death, and prediction of 30-day unplanned readmission may improve outcomes by identifying patients who may benefit from readmission prevention strategies. The goal of this study is to develop machine learning models to stratify risk of 30-day procedure-related mortality and 30-day unplanned readmission in patients undergoing lower extremity infra-inguinal endovascular interventions. We used a cohort of 14,444 cases from the American College of Surgeons National Surgical Quality Improvement Program database. For each outcome, we developed and evaluated multiple machine learning models, including Support Vector Machines, Multilayer Perceptrons, and Gradient Boosting Machines, and selected a random forest as the best-performing model for both outcomes. Our 30-day procedure-related mortality model achieved an AUC of 0.75 (95% CI: 0.71-0.79) and our 30-day unplanned readmission model achieved an AUC of 0.68 (95% CI: 0.67-0.71). Stratification of the test set by race (white and non-white), sex (male and female), and age (≥65 years and <65 years) and subsequent evaluation of demographic parity by AUC shows that both models perform equally well across race, sex, and age groups. We interpret the model globally and locally using Gini impurity and SHapley Additive exPlanations (SHAP). Using the top five predictors for death and mortality, we demonstrate differences in survival for subgroups stratified by these predictors, which underscores the utility of our model.

Interpretable Machine Learning for the Prediction of Amputation Risk Following Lower Extremity Infrainguinal Endovascular Interventions for Peripheral Arterial Disease.

PURPOSE: Severe peripheral artery disease (PAD) may result in lower extremity amputation or require multiple procedures to achieve limb salvage. Current prediction models for major amputation risk have had limited performance at the individual level. We developed an interpretable machine learning model that will allow clinicians to identify patients at risk of amputation and optimize treatment decisions for PAD patients. METHODS: We utilized the American College of Surgeons National Surgical Quality Improvement Program database to collect preoperative clinical and laboratory information on 14,444 patients who underwent lower extremity endovascular procedures for PAD from 2011 to 2018. Using data from 2011 to 2017 for training and data from 2018 for testing, we developed a machine learning model to predict 30 day amputation in this patient population. We present performance metrics overall and stratified by race, sex, and age. We also demonstrate model interpretability using Gini importance and SHapley Additive exPlanations. RESULTS: A random forest machine learning model achieved an area under the receiver-operator curve (AU-ROC) of 0.81. The most important features of the model were elective surgery designation, claudication, open wound/wound infection, white blood cell count, and albumin. The model performed equally well on white and non-white patients (Delong p-value = 0.189), males and females (Delong p-value = 0.572), and patients under age 65 and patients age 65 and older (Delong p-value = 0.704). CONCLUSION: We present a machine learning model that predicts 30 day major amputation events in PAD patients undergoing lower extremity endovascular procedures. This model can optimize clinical decision-making for patients with PAD.

Long-term assessment of Asenapine vs. Olanzapine in patients with schizophrenia or schizoaffective disorder.

INTRODUCTION: We conducted a double-blind 1-year trial of asenapine in patients with schizophrenia or schizoaffective disorder. METHODS: Patients were randomized to asenapine (5 or 10 mg BID; n=913) or olanzapine (10-20 mg QD; n=312), and monitored regularly. RESULTS: Trial completion rates were 38% with asenapine, 57% with olanzapine; main reasons for discontinuation were withdrawal of consent (22%, 16%) and insufficient response (25%, 14%); fewer discontinuations were due to adverse events (6%, 7%). Mean weight gain was 0.9 kg with asenapine, 4.2 kg with olanzapine. Extrapyramidal symptoms reported as adverse events were more common with asenapine. Mean reductions in PANSS total score with asenapine and olanzapine were -21.0 and -27.5 ( P<0.0001); the exclusion of patients who had previous poor experience with olanzapine may have biased the results in favor of olanzapine. Scores on the subjective well-being on neuroleptics scale and functionality measures were similar between groups. CONCLUSION: Asenapine was well tolerated over 1 year of treatment, causing less weight gain than olanzapine but more frequent extrapyramidal symptoms. PANSS total score improved with both agents; the improvement was greater with olanzapine than with asenapine using last observations carried forward but not in an observed-case analysis.

The pharmacokinetics of etodolac in serum and synovial fluid of patients with arthritis.

The pharmacokinetics of etodolac have been evaluated in five patients with arthritis given 200 mg etodolac, twice daily, at 12-hour intervals, for 7 days. Albumin and total protein concentrations were markedly lower in synovial fluid than in serum, and etodolac free fraction was significantly higher. Etodolac readily penetrated into the synovial fluid, and in the postdistributive phase the concentration of free etodolac (i.e., the drug responsible for pharmacologic activity) remained higher than that in serum at all times. No differences in the half-life of etodolac elimination were noted.

Efficacy of mirtazapine for prevention of depressive relapse: a placebo-controlled double-blind trial of recently remitted high-risk patients.

BACKGROUND: The necessity of antidepressant continuation-phase therapy following acute-phase response has resulted in the need to characterize the longer-term efficacy and safety of all new medications. Previous studies using "extension" protocols suggest that mirtazapine has sustained antidepressant effects. The current study was performed to evaluate the efficacy and safety of up to 1 year of mirtazapine therapy, using a more rigorous, randomized, placebo-controlled discontinuation design. METHOD: An intent-to-treat sample of 410 patients meeting DSM-IV criteria for moderate-to-severe recurrent or chronic major depressive episodes began 8 to 12 weeks of open-label therapy with mirtazapine (flexibly titrated, 15-45 mg/day). Thereafter, 156 fully remitted patients (according to Hamilton Rating Scale for Depression and Clinical Global Impressions-Improvement scores) were randomly assigned to receive 40 weeks of double-blind continuation-phase therapy with either mirtazapine or placebo. RESULTS: Mirtazapine therapy reduced the rate of depressive relapse by more than half, with 43.8% of patients relapsing on treatment with placebo as compared with 19.7% of the mirtazapine-treated patients. The discontinuation rate due to adverse events was 11.8% for active mirtazapine therapy versus 2.5% for placebo. Although weight gain was significantly greater in the group receiving active medication during the double-blind phase (p = .001), patients taking mirtazapine gained only 1.4 kg (3.1 lb) across the 40 weeks of continuation therapy, and there was no difference in the rates of weight gain as a newonset adverse event. CONCLUSION: Continuation-phase therapy with mirtazapine is effective and well tolerated.

Development of new antidepressants.

A large number of novel antidepressants acting on a variety of neurotransmitter receptors are currently undergoing clinical evaluation. Most agents have a dual mechanism of action on two or more neurotransmitter receptors, including two serotonin receptors, two noradrenergic receptors, or a combination of serotonin and noradrenergic mechanisms. The most recently approved agent, mirtazapine, is an example of this approach of simultaneously targeting both the serotonergic and noradrenergic systems. Specifically, mirtazapine's alpha 2 antagonism disinhibits both serotonin and norepinephrine neurotransmission while its serotonin-2 and serotonin-3 antagonist properties reduce the side effects normally associated with nonselective serotonin receptor activation by serotonin selective reuptake inhibitors (SSRIs). This approach of "designer polypharmacy" applies principles of rational pharmacologic combinations to enhance efficacy and improve tolerability of the new and emerging antidepressants.

A double-blind, placebo-controlled study comparing mianserin and amitriptyline in moderately depressed outpatients.

We report on the results of a study comparing mianserin with amitriptyline and placebo, in outpatients with major depression (DSM-III 296.2 or 296.3). One hundred and forty-nine patients were randomized to mianserin (n = 50), amitriptyline (n = 50) or placebo (n = 49). Medication was taken in a nightly (qhs) dose. During Week 1, the maximum dose was 60 mg mianserin, 120 mg amitriptyline or two placebo capsules. Beginning at Day 7 (through Day 42) maximum dosages were 150 mg mianserin, 300 mg amitriptyline or five placebo capsules. At multiple weeks and endpoint, statistically significant reductions in the Hamilton Depression Scale (HAM-D) 17- and 21-item scores were recorded for both active drugs compared with placebo. Positive results with the HAM-D were corroborated by other measures of efficacy. There were no statistically significant differences between mianserin and amitriptyline in terms of efficacy; however, the results do suggest a more rapid therapeutic response for mianserin compared with amitriptyline, in terms of percentage of patients showing > or = 50% improvement at Weeks 2 (30% vs 23%) and 4 (61% vs 44%). The most common adverse experiences were somnolence (amitriptyline and mianserin 60%, placebo 31%) and dry mouth (amitriptyline 76%, mianserin 30% and placebo 20%). Our results indicate that mianserin is clearly superior to placebo, compares favorably with amitriptyline, and is a safe, well-tolerated, effective medication in the treatment of depressed outpatients.

Breakdown of articular cartilage proteoglycans by lymphokine-activated macrophages.

Lymphokine supernatants (LE) prepared from antigen sensitive lymphocytes caused an inhibition of migration of macrophages from capillary tubes. Control supernatants (LC) had no effect. The lymphokine supernatants, when added to macrophage cultures (the equivalent of 60 x 10(6) lymphocytes added to 40 x 10(6) macrophages), activated the macrophages so that they secreted the enzyme collagenase after 48 h and 72 h of culture. No collagenase was detected before 48 h or from macrophage supernatants to which LC was added. The macrophage supernatants (LE but not LC) also contained factors (probably enzymes) that, when added to a piece of articular cartilage in medium, caused a partial loss of the hexosamine content of the articular cartilage. These changes were seen as early as after 24 h of culture. Activated macrophages therefore release enzymes that can completely destroy cartilage. Both collagenase and a proteoglycan-hydrolyzing enzyme are released which in vivo might be responsible for the cartilage damage that is found in diseases such as rheumatoid arthritis.

Lymphokine-induced uptake of [14C]glucosamine, [14C]glucose, and [3H]leucine by macrophages.

Lymphokine-activated (LK+) and control (LK-) macrophages were cultured for 66 h and then pulsed with [14C]glucosamine. Uptake of [14C]glucosamine was greater in LK+ than in LK- cultures. If, after 66 h, the medium was replaced with fresh medium and then pulsed with either [14C]glucose or [14C]glucosamine, the uptake of isotope was greatly reduced compared to cultures with no change of medium. However, uptake of both radiolabeled substances was still found to be greater in LK+ cultures than in LK- cultures. Although uptake of both substances was enhanced by lymphokines, the uptake kinetics of each isotope was different. Under similar conditions the uptake of [3H]leucine was not enhanced by lymphokine activation. These data are interpreted to mean that LK+ macrophages are metabolically stimulated and utilize more glucose and glucosamine. The difference in kinetics implies a different utilization by macrophages for each substance.

Unique biochemical and biological features of cathepsin D in rodent lymphoid tissues.

Cathepsin D, an enzyme consistently found to be lysosomal in many cells, has an unusual localization in rat thoracic duct lymphocytes (TDL). After fractionation of homogenates of rat TDL, most of the enzyme activity, as measured at pH 3.6 on denatured bovine hemoglobin, is distributed differently from the other lysosomal enzymes. The enzyme also has some unique properties: it is not inhibited by an antiserum inhibitory for rat liver cathepsin D; it exists in two molecular weight forms (approximately 45,000 and approximately 95,000) both of which have a higher specific activity than rat liver cathepsin D, as determined by studies using the irreversible inhibitor, sodium pepstatin; the high molecular weight form converts to the low molecular weight form after treatment with beta-mercaptoethanol without any loss in activity. These enzymes appear to be restricted to rodent lymphoid tissues. Reasons for considering them to be a type of cathepsin D are given in the text.

Pharmacokinetics of mirtazapine from orally administered tablets: influence of a high-fat meal.

The effect of a high-fat meal on the pharmacokinetics of mirtazapine was studied in 19 healthy normal young male volunteers. In a randomized two-period crossover study, each volunteer received an oral dose of 15 mg of mirtazapine in the form of tablets, in the fasting state and after a high-fat meal, with a washout period of 14 days between the two doses. Serial blood samples were taken and pharmacokinetic parameters calculated and statistically analyzed from mirtazapine plasma levels. The extent of absorption of mirtazapine, as measured by the area under the plasma level versus time curve, was found to be equivalent for the fasting and the fed state. Food intake was shown to have no influence on the elimination of mirtazapine, as measured by its elimination half-life. The rate of mirtazapine absorption, as measured by the peak level (Cmax), was not altered by food. The peak time (tmax), however, in subjects in the fed state showed an increase: the 90%-confidence interval for the median difference ranged from 0.25 to 1.25 h. This was the only effect of food found in this study. It is considered to be of no clinical consequence.

Effect of various pharmacological agents on prostaglandin induced increases in microvascular permeability.

The anti-histamines, pyrilamine and chlorpheniramine, the serotonin antagonist, methysergide, the beta-adrenolytic, phenoxybenzamine and the anti-cholinergic, atropine, antagonized the prostaglandin E1 induced increases in vascular permeability. None of the non-steroidal anti-inflammatory agents acted as antagonists at these doses.

Mirtazapine vs. amitriptyline vs. placebo in the treatment of major depressive disorder.

Patients (n = 150) were randomized to a 6-week, double-blind study to evaluate the relative efficacy and safety of mirtazapine, amitriptyline, and placebo in the treatment of major depressive disorder symptoms. Average daily modal doses were mirtazapine, 18 mg; amitriptyline, 111 mg; and placebo, 4.6 capsules. Mirtazapine- and amitriptyline-treated patients had statistically significantly greater mean Hamilton Rating Scale for Depression (HAM-D) score reductions (weekly visits 1, 2, 4, and endpoint) compared to placebo. These findings were supported by the Montgomery-Asberg Depression Rating Scale (MADRS); the Zung Self-rating Depression Scale (SDS); and the Clinical Global Impressions (CGI) scales. Somnolence and weight gain were the only adverse clinical experiences (ACEs) reported substantially more often by mirtazapine-treated patients than by those in the placebo group. However, more amitriptyline-treated patients reported decreased visual accommodation, dry mouth, dyspepsia, constipation, tachycardia, hypertension, hypotension, discoordination, dizziness, and tremor than mirtazapine- or placebo-treated patients. Results of this study indicate that mirtazapine is more effective than placebo in the treatment of these patients, and superior to amitriptyline in respect to anticholinergic and cardiovascular effects.

Etodolac compared with aspirin: an endoscopic study of the gastrointestinal tracts of normal volunteers.

Forty-eight normal men participated in a 14-day, single blind, single center, multiple dose study of gastric irritation using endoscopy. Subjects were randomly assigned to one of 4 treatment groups after a one week lead-in period, as follows: etodolac 200 mg BID, 400 mg BID, 600 mg BID, or aspirin 975 mg QID. Etodolac at all dose levels produced significantly (p less than or equal to 0.0001) less gastrointestinal irritation than aspirin as assessed by endoscopic examination of the gastric and duodenal sites. There were no significant differences among the 3 etodolac groups.

Destruction of articular cartilage by arthritic synovium in vitro: mechanism of breakdown and effect of indomethacin and prednisolone.

Arthritic synovial tissue when cultured with normal articular cartilage induces a breakdown of articular cartilage proteoglycans. No collagen breakdown occurs unless the macroglobulins are inactivated by pretreatment with acid. Proteoglycan breakdown is most likely due to the continuous secretion of hydrolases by the synovium. A proteolytic enzyme has been found in the culture medium which has a pH optimum around 7.6, and is Ca++ dependent. Both indomethacin and hydrocortisone are inhibitors of proteoglycan breakdown.

The pharmacological properties of fenbufen. A review.

gamma-Oxo(1,1'-biphenyl)-4-butanoic acid (fenbufen) was shown to be an orally and parenterally effective nonsteroidal antiinflammatory, analgesic and antipyretic agent in a variety of animal species. Like other clinically active antiinflammatory drugs such as acetylsalicylic acid (ASA), indometacin and phenylbutazone, fenbufen has demonstrated potent activity in a variety of laboratory test systems including carageenin edema (rats), UV erythema (guinea pigs), adjuvant arthritis (rats), urate synovitis (dogs), phenylquinone writhing (mice), and yeast-induced pyresis (rats). In general, fenbufen was less potent than indomethacin and more potent than ASA, and appeared of special interest because of its high analgetic efficacy and long duration of antiinflammatory and analgetic action. While shown to have ulcerogenic potential in rats at toxic doses, fenbufen was less potent than indometacin in this respect and had a superior margin of gastrointestinal safety in treatment of dogs with urate synovitis. One of fenbufen's major metabolites, 4-biphenylacetic acid (BPAA), was found to be potent inhibitor of prostaglandin (PG) synthesis both in vitro and in vivo with a variety of tissues tested. Fenbufen itself was devoid of this anti-PG-synthetase activity although it interacted with prostaglandins in other ways. These results, coupled with the fact that only BPAA showed pharmacological activities when applied locally, led to the conclusion that BPAA was the principle responsible for fenbufen's antiinflammatory action. Fenbufen thus appears to be a pro-drug capable of circumventing at least some of the gastric toxicity usually incurred when compounds, which are themselves capable of inhibiting PG synthesis, are introduced directly into the stomach. Fenbufen's relatively low gastric toxicity in dogs and humans seems to substantiate this hypothesis. The pharmacological evidence indicates that fenbufen should be a highly effective and clinically useful antiinflammatory, analgetic and antipyretic drug.