A process mining- deep learning approach to predict survival in a cohort of hospitalized COVID-19 patients.

BACKGROUND: Various machine learning and artificial intelligence methods have been used to predict outcomes of hospitalized COVID-19 patients. However, process mining has not yet been used for COVID-19 prediction. We developed a process mining/deep learning approach to predict mortality among COVID-19 patients and updated the prediction in 6-h intervals during the first 72 h after hospital admission. METHODS: The process mining/deep learning model produced temporal information related to the variables and incorporated demographic and clinical data to predict mortality. The mortality prediction was updated in 6-h intervals during the first 72 h after hospital admission. Moreover, the performance of the model was compared with published and self-developed traditional machine learning models that did not use time as a variable. The performance was compared using the Area Under the Receiver Operator Curve (AUROC), accuracy, sensitivity, and specificity. RESULTS: The proposed process mining/deep learning model outperformed the comparison models in almost all time intervals with a robust AUROC above 80% on a dataset that was imbalanced. CONCLUSIONS: Our proposed process mining/deep learning model performed significantly better than commonly used machine learning approaches that ignore time information. Thus, time information should be incorporated in models to predict outcomes more accurately.

[Infant mortality in Germany (2008-2012)--lower in the former German Democratic Republic?]. / Säuglingssterblichkeit in Deutschland (2008-2012)--niedriger im Osten?

BACKGROUND: German infant mortality is ranked near the median of European countries. In Germany infant mortality is significantly higher in the German Federal Republic compared with the former German Democratic Republic. This is often used as reason for a call for structural requirements and minimum caseload for the care for very low birth weight infants. METHOD: Neonatal and infant mortality were calculated for the 16 German federal states with data from the German statistical federal office for the years 2008-2012. RESULTS: Considerable variations were found for the neonatal (1.34-3.61‰, total Germany 2.31‰) and the infant (2.38-5.20‰, 3.47‰) mortality. The rate of stillborn infants was 3.56‰. A lower neonatal mortality in the former German Democratic Republic (1.62‰ vs. 2.44‰, p<0.0001, Chi-squared test) could not be confirmed for preterm infants with birth weight less than 1 500 g. In the former German Democratic Republic stillbirth was significantly more frequent in preterm infants with birth weight 500-999 g (p<0.0001). Combined stillbirth and neonatal mortality showed no difference between the German Federal Republic and former German Democratic Republic (5.45‰ and 5.29‰, respectively, n.s.; infants less than 500 g birth weight were excluded). The average number of preterm infants per perinatal centre and federal state had no influence on state specific neonatal mortality. CONCLUSION: If stillborn infants were accounted for no difference was found between the German Federal Republic and the former German Democratic Republic regarding mortality. Comparing infant mortality of different countries has to account for stillborn infants. Considerable variation of neonatal mortality is persisting throughout Germany despite structural requirements and introduction of a minimum caseload since 2005. A lower infant mortality in the former German Democratic Republic and implications drawn from are not supported by the presented nationwide data from the German statistical federal office.

High shear mixer granulation using food grade binders with different thickening power.

Mixer agglomeration, in particular high shear wet granulation, is a unit operation typically used in the pharmaceutical industry to improve the flowability, the compressibility, the dosing accuracy during tableting or the content uniformity of a blend. Thanks to its advantages (production of spherical and dense granules, reduction of production time), this technique can be potentially successful also in the food industry as for example in the production of dietary supplements. In this work four thickening agents (povidone, maltodextrin, k-carrageenan and xanthan gum) have been tested to study their effects on the granule growth behavior and on some technologically relevant granule properties (size, shape, strength and flowability). Experiments highlighted the full feasibility of the process and the possibility of using these agents to get products with satisfactory technological properties. The dependence of product properties on the formulation variables (water and binder amount) has been analyzed according to a multivariate approach and a robust predictive tool for the granule size has been developed. Furthermore it was observed that a reduced amount of binding liquid (water) can be used in the presence of strongly thickening binders with a reduction up to 25%. This would decrease drying time and energy requirement and be beneficial especially in the food and food supply industry where products have generally lower added value than in the pharmaceutical one and reducing production costs is critical.

Computed tomography (CT) colonography with CT arteriography and venography for the workup of intestinal transplant candidates.

Prior to intestinal transplantation, prospective candidates must undergo a series of radiologic examinations to address a variety of clinical issues. To date, little literature exists to guide physicians in this preoperative assessment. Multiple imaging studies can provide overlapping information. We have developed a simple two- or three-test protocol to streamline the workup. Sixteen adult patients presented as potential intestinal transplant candidates to Georgetown University Hospital. All but two patients underwent the full protocol of a biphasic IV contrast-enhanced computed tomography (CT) scan of the chest, abdomen, and pelvis with rectal carbon dioxide, an upper gastrointestinal study with small bowel follow through, and fistulogram when appropriate. Three-dimensional (3-D) reconstructions of the vascular anatomy as well as the colon were also generated. A telephone survey to other transplant centers was additionally conducted to compare radiographic evaluations. Overall, 15 of the 16 scans were diagnostic. One patient required a barium enema. Mean examinations per patient was 2.4. Only one of seven other centers was performing CT colonography in prospective intestinal transplant candidates. Our protocol provided all the necessary anatomic information needed to evaluate prospective transplant candidates. CT colonography with angiography is a suitable alternative to more time-consuming radiological studies.

Calcium and phosphor intake in preterm infants: sensitivity and specifity of 6-hour urine samples to detect deficiency.

Aim of the present study was to test whether six-hour (6 h) urine specimens predict the 24-hour (24 h) mineral homeostasis in individual infants born preterm. Urinary Calcium (Ca) and Phosphate (P) concentrations were studied in 60 stable infants; gestational age 34 (25-42) weeks. In 58 infants four 6 h urine specimens and in 2 infants all spot urine specimens obtained within 24 h were analyzed. In 39 infants born preterm coefficients of variation were 0.42 (SD 0.26) and 0.41 (SD 0.26) for Ca and P measurements in the four 6 h urine specimens obtained within 24 h, respectively, The mineral homeostasis of the infants was defined as Ca or P surplus homeostasis if the 24 h urinary concentrations were ≥1 mmol/l. The sensitivity, specificity, and PPV of a 6 h urinary specimen to predict Ca deficiency homeostasis (24 h urinary Ca <1 mmol/l) were 0.93 (0.77-0.98; 95%CI), 0.72 (0.43-0.90) and 0.90 (0.74-0.96). The sensitivity, specificity and PPV for urinary P were 0.8 (0.38-0.96), 0.97 (0.85-0.995), and 0.8 (0.38-0.96). In conclusion, in infants born preterm on regular 3 or 4 h feedings, 6 h urine sampling is sufficiently precise for prediction of Ca and P mineral deficiency homeostasis (PPV 0.92 and 0.83). However, measurements at regular intervals (twice weekly) are recommended not to miss any infant in mineral deficiency homeostasis.

[Current outcome quality in the care of preterm infants with birth weight less than 1 500 g as a basis for regionalisation of risk pregnancies]. / Aktuelle Ergebnisqualität der Versorgung von Frühgeborenen <1 500 g Geburtsgewicht als Grundlage für eine Regionalisierung der Risikogeburten.

BACKGROUND: For preterm infants an association between patient volume and mortality has been described. METHODS: Outcome variables were evaluated for 28 hospitals in Baden-Württemberg for the years 2004-2008. Hospitals with high patient volume were compared to hospitals with a lower patient volume. RESULTS: Outcomes for 1 164 infants in 2008 and for 4 775 infants in 2004-2008 were analysed. In 2008, mortality of preterm infants less than 32 weeks gestational age (GA) was 9.2% (n=402) in the 5 major hospitals compared to 6.5% (n=520) in the other hospitals (combined mortality 7.7%, n. s., chi-square test). In the years 2004-2008, mortality showed a greater variation in hospitals with a patient volume below 50 and mean mortality was 21.1% higher for infants less than 500 g BW. Hospitals with a patient volume >or= 50 had a lower mortality for infants with BW below 500 g and between 500 g and 749 g (18% and 11%, chi-square test: p<0.05 and <0.01, respectively). For preterm infants with GA below 24 weeks and between 24 and 25 weeks, patient volume and mortality were negatively correlated (p<0.01 and <0.0001, respectively). For infants with a BW >or= 750 g or a GA >or= 26 weeks patient volume had no effect on outcome. CONCLUSION: Regionalisation of preterm infants with BW less than 750 g and a GA less than 26 weeks may contribute to reduce mortality. Infants with BW >or= 750 g and a GA >or= 26 weeks may not benefit from indirect quality indicators such as patient volume.

[Thoracic alterations after surgical closure of the ductus arteriosus botallo in preterm infants]. / Thoraxveränderungen nach operativem Ductus-arteriosus-Botallo-Verschluss bei Frühgeborenen.

BACKGROUND: Left-sided thoracotomy for ligation of patent ductus arteriosus (PDA) dissects the musculus latissimus dorsi and notches a small part of the musculus trapezius. After ductal closure the 4 (th) and 5 (th) rib are adapted. This follow-up study investigated if mid- or long-term consequences on the thorax occur after this procedure. PATIENTS AND METHODS: Status of the thoracic scar, functionality of the shoulder and presence of scapulata alata or scoliosis was evaluated at median age of 6 years (range: 2.9-11.9) in 57 pre-term infants (30 male; gestational age 26 weeks (24-32); birth weight 805 g (450-2140)). RESULTS: Scoliosis was diagnosed in 1 patient (=1.8%) with Rubinstein-Taybi syndrome. The length of the thoracic scar (13.8 cm; 9.4-25.5) correlated with the patient's age (r=0.61; p=0.001). The scar was relocatable except for one case. The distance of the ventral end of the scar to the nipple was 2 cm or less in 22% of the female patients. None of the patients showed impaired function of the shoulder. Scapula alata was found in 16 (28%) patients. CONCLUSION: Thoracotomy for PDA ligation was not associated with an increased risk for scoliosis or disturbed function of the shoulder. One quarter of all infants developed scapula alata which meant an aesthetic issue for some parents.

Effect of oestradiol and progesterone replacement on bronchopulmonary dysplasia in extremely preterm infants.

OBJECTIVE: To study whether postnatal replacement of oestradiol and progesterone may help to prevent bronchopulmonary dysplasia (BPD). METHODS: This randomised placebo-controlled double-blind study enrolled 83 infants of <29 weeks gestational age and 1000 g birth weight requiring mechanical ventilation within 12 h after birth. Oestradiol (2.5 mg/kg/day) and progesterone (22.5 mg/kg/day) were given by continuous intravenous infusion of a standard lipid emulsion (15 ml/kg/day) in the replacement group (ESTRA-PRO). The placebo group received the same lipid emulsion without oestradiol or progesterone. A replacement period of at least 2 weeks but not >4 weeks was strived for and defined as "according to protocol". The primary outcome variable was the incidence of BPD or death. RESULTS: The median birth weight was 670 g (min-max 400-990 g) and the gestational age 25 weeks (23.1-28.1 weeks). The incidence of BPD or death was 48% in the placebo group and 44% in the ESTRA-PRO group (p = 0.38, one-sided testing, intention to treat analysis). In infants treated according to protocol, 32% (9 of 28) in the placebo group and 14% (3 of 21) in the ESTRA-PRO group developed BPD (p = 0.08). CONCLUSION: Replacement of oestradiol and progesterone was not effective for prevention of BPD or death in extremely preterm born infants. Better-powered trials are needed to evaluate this new approach.

Late response to radiochemotherapy in pediatric glioblastoma: report on two patients treated according to HIT-GBM protocols.

High-grade gliomas in children are rare and the best treatment is undetermined. The German language group study HIT-GBM compares various induction protocols for subsequent patient cohorts. Currently, cisplatinum, etoposide, ifosfamide, and vincristine are given simultaneously with extended-field radiotherapy. Imaging is done 3 weeks after to define treatment response, followed by 6-weekly controls during consolidation with lomustine, vincristine, and prednisone. The authors report on 2 patients with incompletely resected glioblastoma multiforme in which response was lacking 3 weeks after radiochemotherapy but became evident 12 weeks later. This suggests that later time points are required to assess induction protocol response.

Calcium and phosphorus retention in extremely preterm infants supplemented individually.

UNLABELLED: The aim of this study was to investigate the correlation between the retention of calcium (Ca) and phosphorus (P) and weight gain and intake of Ca and P when using the concept of individualized Ca and P supplementation in extremely low birthweight infants. Three-day Ca and P balances were performed in 20 infants with a mean gestational age of 26.6wk (between 24.1 and 28.7 wk) and a birthweight of 744 g (450-990), when the infant was able to tolerate at least 100 ml/kg/d of milk. The daily supplementation with Ca and P was individually adjusted to achieve a simultaneous excretion of > or = 1.2 mmol/L Ca and > or = 0.4 mmol/L P in the urine. In 16 of the 20 infants, the urinary concentrations of both Ca and P exceeded the lower limits. The retention of Ca (mean 3.8 mmol/kg/d, minimum 0.9; maximum 8.1; 57% of intake, 34-80) and P (2.4,1.1-4.2; 76%, 52-96) was significantly correlated with both the daily weight gain (16 g, 3-28; Ca r2 = 0.22, p = 0.02; P r2 = 0.21, p = 0.03) and the intake of Ca (6.5 mmol/kg/d, 2.4-10.2; r2 = 0.67, p < 0.001) and P (3.1, 1.9-5.3; r2 = 0.85, p < 0.0001). The molar ratio of the Ca and P intake was 2.2 (1.3-4.0). CONCLUSION: It was found that Ca and P retention was a function of growth and intake.

Epinephrine treatment of hypotension in very low birthweight infants.

UNLABELLED: The aim of this study was to examine the influence of a continuous infusion of epinephrine (adrenaline) on mean arterial blood pressure (MABP), heart rate, urine output and base deficit in very low birthweight infants (VLBWI) with systemic hypotension. In VLBWI who received an infusion of epinephrine for at least 12 h the mean urine output, administered fluid volume, base deficit and administered buffer 12 h before and 12 h during the infusion were recorded. If the infusion was shorter, but given for at least 2 h, the mean heart rate and MABP 2 h before and 2 h during the infusion were recorded. Thirty-one infants with a gestational age of 26 (23-30) wk [median (minimum-maximum)] and birthweight 690 (390-1310) g were included in this retrospective chart review. The patients received an infusion of epinephrine at a postnatal age of 3 (1-21) d. The doses ranged between 0.05 and 2.6 microg kg(-1) per minute within the first 24 h of administration. Three of 31 infants received epinephrine on 2 different occasions. The MABP [+7 (-1 to 13) mmHg, p=0.000001] and the heart rate [+10 (-10 to 42) bpm, p=0.000036] increased significantly (n = 34), whereas total volume administration and urine output remained the same between the 2 periods (Wilcoxon matched pairs test). The base deficit increased significantly [-3 (-10.2 to 2.6), p = 0.0014, n = 19] without a change in the administration of buffer. CONCLUSION: The infusion of epinephrine increased the MABP and the heart rate without decreasing urine output in VLBWI with hypotension not responding to a dopamine infusion up to 15 microg kg(-1) per minute. A potential adverse effect was an increase in metabolic acidosis.

Restrictive interatrial communication with protein-losing enteropathy and coagulopathy in hypoplastic left heart syndrome after Norwood palliation.

The Norwood procedure is one option for neonates born with hypoplastic left heart syndrome. We describe a case of an infant with hypoplastic left heart syndrome, palliatively repaired with the Norwood procedure. The infant developed restriction of the interatrial communication, despite atrioseptectomy at the first stage of palliation. Consequently, a protein-losing enteropathy with severe coagulopathy developed which resolved after a repeat atrioseptectomy.

Gender-related plasma levels of progesterone, interleukin-8 and interleukin-10 during and after cardiopulmonary bypass in infants and children.

BACKGROUND: It is known that proinflammatory and anti-inflammatory cytokines are released during and after cardiopulmonary bypass (CPB) in infants and children. Sex steroids are known to have immunomodulatory functions, and release of the anti-inflammatory cytokine IL-10 is stimulated by progesterone in vitro. The purpose of the present study was to investigate the plasma levels of progesterone, IL-8 (proinflammatory cytokine) and IL-10, and to relate them to sex and postoperative morbidity. METHOD: Eighteen infants and children (eight female) undergoing CPB were prospectively studied. Plasma levels of progesterone, IL-8 and IL-10 were determined before and 10 min after the start of CPB, and immediately after CPB; and 6 h, 24 h, 3 days and 7 days postoperatively. Organ dysfunction was identified on the basis of arbitrarily defined criteria. RESULTS: After CPB, all patients showed significant increases in plasma levels of progesterone, IL-8 and IL-10. Plasma levels of IL-10 were significantly higher in female patients, except for during the immediate postoperative period. According to the criteria used, six out of 10 male patients, but none of the female patients developed multiple organ dysfunction (MOD). CONCLUSION: The present study shows that CPB induces a significant and marked increase in plasma levels of progesterone in infants and children. Studies of administration of progesterone-blocking substances to male and female animals may help to elucidate the roles of sex and progesterone in the setting of CBP.

Management of the extremely preterm infant: is the replacement of estradiol and progesterone beneficial?

This review presents data to suggest that postnatal estradiol and progesterone replacement therapy may be beneficial in preterm infants. During pregnancy, maternal plasma levels of estradiol and progesterone increase up to 100-fold compared to the nonpregnant status. The fetus is also exposed to these increasing hormone levels. After delivery, estradiol and progesterone levels drop by a factor of 100 within 1 day. Whereas this is a physiological condition for an infant born at term, preterm delivery means withdrawal from the placental supply of these hormones at an earlier developmental stage. Seventy years ago, the idea was raised that preterm infants may benefit from the replacement of estrogens. Studies in which estrogen was injected subcutaneously showed only a slightly better bodyweight gain compared to placebo-treated controls and therefore routine use was not established. The effective treatment of postmenopausal osteoporosis with hormone replacement therapy led to a pilot study of estradiol and progesterone therapy to prevent osteopenia of prematurity. The highest median bone mineral accretion rate was found in the replacement group when the supplementation with calcium and phosphorus was also sufficient. None of the previous studies dealing with estrogen replacement controlled for achieved plasma levels of estradiol in the infants. In our controlled randomised pilot study with 30 preterm infants (15 in each group), we aimed to maintain intra-uterine plasma levels of estradiol and progesterone. Preterm infants with replacement of estradiol and progesterone for 6 weeks postnatally showed trends to higher bone mineral accumulation. In addition, a trend towards a lower incidence of chronic lung disease was found. Neurodevelopmental follow-up showed normal psychomotor development in infants given estradiol and progesterone, whereas the untreated infants (controls) showed a trend towards delayed development. Recent research emphasises that estradiol and progesterone may be important for brain development. Thus, while there is data indicating that postnatal estradiol and progesterone replacement therapy may be beneficial in preterm infants, experience with this new therapy is limited and extensive research is needed to address the potential benefits and to rule out adverse effects.

Follow-up examination at the age of 15 months of extremely preterm infants after postnatal estradiol and progesterone replacement.

A randomized controlled pilot study was performed with a sample of extremely preterm infants to evaluate the impact of postnatal estradiol and progesterone replacement on postnatal bone mineral accretion. Twenty-five of 30 infants in the pilot study survived, and of these, 24 infants were available for the follow-up examination at a median chronological age of 18.1 months (minimum-maximum, 17.0--20.6) corresponding to a corrected age of 14.8 months (minimum-maximum, 12.9--17.4). Somatic growth data and bone mineralization showed no differences between the hormone-treated and control group infants. The deviation of the skeletal age from the corrected age was 0.0 months (minimum-maximum, -7.7 to 7.4) for hormone-treated infants compared with -1.7 months (minimum-maximum, -7.5 to 5.9) for the control group. The Bayley scales mental and psychomotor developmental indexes were 89 (minimum-maximum, 71--107) and 101 (minimum-maximum, 49--121) for the hormone-treated infants and 93 (minimum-maximum, 49--111) and 71 (minimum-maximum, 49--121) for the control group infants, respectively (mental developmental index, P = 1.0; psychomotor developmental index, P = 0.14). The normal psychomotor development in the hormone-treated infants compared with the below average development in the control group infants is encouraging and indicates the potentially important integrative role of sex steroids for the developing brain. Larger studies on the effects of the postnatal replacement of estradiol and progesterone in extremely preterm infants are warranted.

Unusual electrocardiogram findings in a preterm infant after fetal tachycardia with hydrops fetalis treated with flecainide.

Fetal tachycardia and signs of hydrops fetalis were diagnosed at 29 weeks of gestation. The heart rate normalized by combined treatment with digoxin and flecainide and was followed by improvement of the hydrops. Premature labor led to delivery at 33 weeks of gestation. The newborn infant showed mild respiratory distress and was in a hemodynamically stable condition. Marked QT segment anomalies on the electrocardiogram during the first postnatal days resolved completely within 3 weeks. They were unlikely to be attributable to myocarditis or myocardial infarction. We speculate that these anomalies were caused by the maternal flecainide therapy.

The replacement of oestradiol and progesterone in very premature infants.

The idea of replacing 17beta-oestradiol (E2) and progesterone (P) in preterm infants is based on the observation that during pregnancy E2 and P plasma concentrations rise in the mother and the fetus by a factor of 100. Disruption of the placental supply of these hormones is a physiological event for an infant delivered at term. A preterm infant is deprived from this supply at an earlier developmental stage. In vitro and in vivo data are discussed, and they highlights the potential benefit of E2 and P on the development of different organ systems. The postnatal replacement of E2 and P has the aim of maintaining in utero plasma concentrations. In the first randomized clinical study in 30 extremely preterm infants, E2 and P were replaced postnatally for a total of 6 weeks. With a median intravenous replacement of 8.4 micromol/kg/day of E2 (4.2-22.9) and 67.4 micromol/kg/day of P (35.7-87.0), plasma levels of E2 and P were maintained within the intrauterine reference values of 7.3-22.0 nmol/L and 0.95-1.9 micromol/L, respectively. Three- to sixfold higher dosages were needed via the transepidermal route. Trends towards an improved postnatal bone mineral accretion and a reduced incidence of chronic lung disease were found. Further studies are warranted to clarify the potentially important role of E2 and P for the postnatal development of an extremely preterm infant.