Risk Factors for Postprocedural Arterial Ischemic Stroke in Children With Cardiac Disease.

BACKGROUND AND PURPOSE: Cardiac pathologies are the second most frequent risk factor (RF) in children with arterial ischemic stroke (AIS). This study aimed to analyze RFs for AIS in children with cardiac disease and cardiac intervention. METHODS: Data were drawn from the Swiss Neuropediatric Stroke Registry. Patients with cardiac disease and postprocedural AIS registered from 2000 until 2015 were analyzed for the cause of cardiac disease and for potential RFs. RESULTS: Forty-seven out of 78 children with cardiac disease had a cardiac intervention. Of these, 36 presented a postprocedural AIS. Median time from cardiac intervention to symptom onset was 4 days (interquartile range, 2-8.5); time to diagnosis of AIS was 2 days (interquartile range, 0-5.8). Main RFs for postprocedural AIS were hypotension, prosthetic cardiac material, right-to-left shunt, arrhythmias, low cardiac output, and infections. CONCLUSIONS: In children with postprocedural AIS, time to diagnosis was delayed. Most patients presented multiple potentially modifiable RFs as hemodynamic alterations and infections.

Prevention of Low Cardiac Output Syndrome After Pediatric Cardiac Surgery: A Double-Blind Randomized Clinical Pilot Study Comparing Dobutamine and Milrinone.

OBJECTIVES: Dobutamine and milrinone are commonly used after open-heart surgery to prevent or treat low cardiac output syndrome. We sought to compare efficacy and safety of these drugs in pediatric patients. DESIGN: Prospective, single-center, double-blinded, randomized clinical pilot study. SETTING: Tertiary-care university children's hospital postoperative pediatric cardiac ICU. PATIENTS: After written consent, 50 consecutive patients (age, 0.2-14.2 yr; median, 1.2 yr) undergoing open-heart surgery for congenital malformations were included. INTERVENTIONS: After cardiopulmonary bypass, a continuous infusion of either dobutamine or milrinone was administered for the first 36 postoperative hours. Maximum dose: dobutamine 6 µg/kg/min, milrinone 0.75 µg/kg/min. MEASUREMENTS AND MAIN RESULTS: There were no significant differences in demographic data, complexity of surgery, and intraoperative characteristics between the two study groups (dobutamine vs milrinone). Efficacy was defined as need for additional vasoactive support, which did not differ between groups (dobutamine 61% vs milrinone 67%; p = 0.71). Sodium nitroprusside was used more often in the dobutamine group (42% vs 13%; p = 0.019). Systolic blood pressure showed a trend toward higher values in the dobutamine group, whereas both drugs increased heart rate early postoperatively. Echocardiography demonstrated a consistently good cardiac function in both groups. Central venous oxygen saturation, serum lactate levels, urine output, time to chest tube removal, length of mechanical ventilation, ICU, and hospital stay were similar in both groups. Both drugs were well tolerated, no serious adverse events occurred. CONCLUSIONS: Dobutamine and milrinone are safe, well tolerated, and equally effective in prevention of low cardiac output syndrome after pediatric cardiac surgery. The hemodynamic response of the two drugs is comparable. In uncomplicated cases, a trend toward the more cost-saving dobutamine might be anticipated; however, milrinone demonstrated a trend toward higher efficacy in afterload reduction.

Transthoracic intracardiac catheters in pediatric cardiac patients: A single-center experience.

BACKGROUND: Transthoracic intracardiac catheters are frequently inserted in children during congenital heart surgery for monitoring and vascular access purposes. Their use entails a small potential risk. AIM: We aimed to evaluate both catheter-associated morbidities related to maintenance and removal of transthoracic intracardiac catheters in pediatric cardiac patients, and predictors for catheter-associated adverse events. METHODS: Single-center retrospective cohort study of prospectively collected data of children aged 0-14 years receiving a transthoracic intracardiac catheter inserted in the operating room during 7 consecutive years at the University Children's Hospital Zurich. RESULTS: A total of 115 transthoracic intracardiac catheters were placed in 112 patients: 45 right atrial, 68 left atrial, and 2 pulmonary artery catheters. Five catheters (4.3%) had to be removed due to catheter-associated adverse events (infection, 2; nonfunction, 2; and leakage 1). After catheter removal, 19% of patients suffered adverse events, these were minor in 16 (14%) and serious in 6 (5.1%) (symptomatic bleeding in four (3.5%) patients, pericardial tamponade leading to death in one (0.8%), and tension pneumothorax in one (0.8%)). Catheter position in the right atrium and the need for platelet transfusion prior to removal were risk factors for adverse events. CONCLUSIONS: Transthoracic intracardiac catheters are useful in the management of specific patient groups with complex congenital heart defects. Adverse events do occur; most of them do not require intervention. The insertion technique plays an important role in avoiding adverse events. Strict guidelines for the use and removal of transthoracic intracardiac catheters are required. Low platelet count should delay catheter removal. The wealth of information and therapeutic options offered by these catheters appear to outweigh the associated potential adverse events in this specific patient group.

Successful Lung Transplant After Prolonged Extracorporeal Membrane Oxygenation (ECMO) in a Child With Pulmonary Hypertension: A Case Report.

INTRODUCTION: The use of extracorporeal membrane oxygenation (ECMO) is considered a risk factor for, or even a potential contraindication to, lung transplantation. However, only a few pediatric cases have been described thus far. CASE PRESENTATION: A 9-year-old boy with idiopathic pulmonary arterial hypertension developed cardiac arrest after the insertion of a central catheter. ECMO was used as a bridge to lung transplantation. However, after prolonged resuscitation, he developed medullary ischemia and medullary syndrome. After 6 weeks of ECMO and triple combination therapy for pulmonary hypertension, including continuous intravenous prostacyclin, he was weaned off support, and after 2 weeks, bilateral lung transplantation was performed. At 4 years post-transplant, he has minimal problems. The medullary syndrome has also alleviated. He is now back to school and can walk with aids. CONCLUSIONS: Increasing evidence supports the use of ECMO as a bridge to LT, reporting good outcomes. In the modern era of PAH therapy, it is feasible to use prolonged ECMO support as a bridge to lung transplant, with the aim of weaning off this support; however, its use requires more experience and knowledge of long-term outcomes.

Incidence and predictors of cardiac catheterisation-related arterial thrombosis in children.

OBJECTIVE: Arterial thrombosis is one of the most reported complications of cardiac catheterisation (CC) in children. The aim of the study was to evaluate the incidence and predictors of arterial thrombosis in children with cardiac diseases (CDs). METHODS: During 12 consecutive months, all children aged 0-19 years undergoing CC of the femoral arteries were included in this observational study. After CC, clinical evaluation of impaired limb perfusion was performed according to local guidelines. Doppler ultrasonography was performed when decreased limb perfusion was suspected. RESULTS: 123 children (30% aged <12 months, 70% aged >12 months) underwent CC. Arterial thrombosis occurred in 14 of the 123 children (11.4%). Twelve cases (12/14=86%) of arterial thrombosis occurred in infants aged <12 months and 2 (2/14=14%) in older children. Overall younger age (p<0.01, OR (95% CI) 0.49 (0.28 to 0.86)) and low body weight (p<0.004, OR (95% CI) 0.78 (0.65 to 0.92)) were significantly associated with an increased risk of arterial thrombosis. Cyanotic CD (p=0.07, OR (95% CI) 2.87 (0.90 to 9.15)) showed a trend towards increased thrombotic risk. CONCLUSIONS: Arterial thrombosis is a common complication of CC in infants. Diagnosis of CC-related arterial thrombosis remains a challenge. Well-defined clinical monitoring protocols may be valuable methods for timely detection and treatment of arterial thrombosis.

Role of sevoflurane in organ protection during cardiac surgery in children: a randomized controlled trial.

OBJECTIVES: The protective effects of volatile anaesthetics against ischaemia-reperfusion injury have been shown in vitro, but clinical studies have yielded variable results. We hypothesized that, in children, sevoflurane provides superior cardioprotection after cardiac surgery on cardiopulmonary bypass (CPB) compared with totally intravenous anaesthesia (TIVA). METHODS: In this randomized controlled, single-centre study, 60 children with cyanotic and acyanotic heart defects undergoing elective cardiac surgery under CPB (RACHS-1 1-3) were randomized to sevoflurane or TIVA (midazolam <6 months of age, propofol >6 months of age). The primary end-point was the postoperative peak cardiac troponin I/T (cTnI/T). Perioperative cardiac function (as determined by brain-type natriuretic peptide, echocardiography and postoperative vasopressor/inotrope requirements), short-term clinical outcomes (duration of intubation, intensive care unit and hospital length of stay), postoperative inflammatory profile, and pulmonary, renal and liver function were defined as secondary end-points. Analysis of variance was used for statistical analysis. RESULTS: There was no statistically significant difference in postoperative peak troponin values or any of the secondary end-points. In the subgroup of acyanotic patients under 6 months, sevoflurane led to significantly lower postoperative troponin levels compared with midazolam [reduction of 54% (95% confidence interval 29-71%, P = 0.002)], without any differences in secondary outcome parameters. CONCLUSIONS: Sevoflurane did not provide superior myocardial protection in our general paediatric cardiac surgical population. In children under 6 months, however, sevoflurane might be beneficial in comparison with midazolam. The conditioning effects of sevoflurane in specific paediatric subgroups need to be further investigated.

Postoperative arrhythmia in patients with bidirectional cavopulmonary anastomosis.

OBJECTIVES: Bidirectional cavopulmonary anastomosis (BDCPA) is part of the staged Fontan palliation for congenital heart defects with single-ventricle morphology. The aim of this study was to describe incidence and characteristics of early postoperative arrhythmias in patients undergoing BDCPA. METHODS: Retrospective analysis of 60 patients undergoing BDCPA at the age of <12 months from 2001 to 2008 at a single centre. Arrhythmias were subclassified in sinus bradycardia, premature atrial/ventricular contraction, supraventricular tachycardia and atrioventricular block. The groups were compared according to age at operation and diagnosis. Postoperative follow-up data were included until Fontan completion. RESULTS: Postoperative arrhythmia was observed in 20 patients: 12 temporary and 8 persisting until hospital discharge. Sinus bradycardia is a common postoperative arrhythmia and occurred in 16 patients (9 transient, 7 persistent until hospital discharge). One patient undergoing BDCPA and a Damus-Kaye-Stansel procedure had a persisting first-degree atrioventricular block. The occurrence of a postoperative arrhythmia was independent of age and diagnostic group (hypoplastic left heart vs non-hypoplasic left heart). After hospital discharge, five of the eight arrhythmia resolved spontaneously resulting in 2 patients with sinus bradycardia and 1 patient with a first-degree AV block immediately before the Fontan completion was undertaken. CONCLUSIONS: Postoperative arrhythmias in patients with BDCPA occur early after surgery and are temporary. Severe and life-threatening arrhythmias are rare although the interventions are complex and the patients very young. The most common arrhythmia is sinus bradycardia.

Mortality and neurodevelopmental outcome at 1 year of age comparing hybrid and Norwood procedures.

OBJECTIVES: Neonates with hypoplastic left heart syndrome (HLHS) are at risk of high mortality and neurodevelopmental morbidity. As an alternative to Norwood-type stage I palliation, the hybrid procedure has been developed. It consists of bilateral pulmonary artery banding, catheter-based stenting of the arterial duct and balloon atrioseptostomy and delays open-heart surgery. Thus, it may be associated with a better outcome. The aim of this study was to determine the mortality and neurodevelopmental outcome in patients with HLHS and other univentricular heart (UVH) defects treated with hybrid or Norwood procedures. METHODS: Thirty-one children (18 males) with HLHS and other UVH defects undergoing Norwood or hybrid procedure between 2004 and 2008 were consecutively enrolled. Mortality and neurodevelopmental outcome at 1 year of age were determined. RESULTS: One-year mortality was 36% (31% in the hybrid vs. 39% in the Norwood group, P=0.71). Predictors of mortality were lower birth weight (P=0.02), older age at first procedure (P=0.02) and smaller size of ascending aorta (P=0.05). Overall, median psychomotor development index (PDI) and mental development index (MDI) of the Bayley Scales of Infant Development II were lower than the norm of 100 [PDI 57 (49-99), P<0.001; MDI 91 (65-109), P=0.002]. No effect of surgical treatment on neurodevelopmental outcome was found. Predictors of impaired motor outcome were length of hospital stay (LOHS) (P=0.01), lower body weight at second procedure (P=0.004) and female sex (P=0.01). Predictors of impaired cognitive outcome were longer mechanical ventilation time (P=0.03), intensive care unit stay (P=0.04) and LOHS (P<0.001), respectively. CONCLUSIONS: Mortality at 1 year of age is comparable between patients undergoing hybrid and Norwood procedures. Early neurodevelopmental outcome is significantly impaired in patients with both HLHS and other UVH defects. Multicentre randomized studies are needed to determine the long-term neurodevelopmental outcome of children treated with the hybrid procedure.

Do predictors exist for a successful withdrawal of preoperative prostaglandin E(1) from neonates with d-transposition of the great arteries and intact ventricular septum?

Prostaglandin E(1) (PGE(1)) is given to neonates with d-transposition of the great arteries (d-TGA) to reduce cyanosis by reopening and maintaining the patency of the ductus arteriosus. To avoid side effects, this medication can be stopped for hemodynamically stable patients after balloon atrial septostomy (BAS). A consecutive series of neonates with d-TGA and an intact ventricular septum (IVS) presenting from 2000 through 2005 was analyzed retrospectively to search for side effects of PGE(1) and to identify predictors for a safe preoperative withdrawal. The medication was stopped for hemodynamically stable patients with transcutaneous oxygen saturations higher than 80% after BAS and reinitiated for patients with an oxygen saturation lower than 65%. Patients successfully weaned were compared with those who had failed weaning in terms of atrial septal defect (ASD) size, ductus arteriosus size, and the transcutaneous oxygen saturation. Prostaglandin E(1) was initiated for all 43 neonates with d-TGA. The median maintenance dose of PGE(1) was 0.00625 µg/kg/min (range, 0.00313-0.050 µg/kg/min) for a median duration of 6 days (range, 1-12 days). For 16 patients, PGE(1) was preoperatively withdrawn but then had to be reinitiated for 7 of the 16 patients. No predictors for a successful weaning of PGE(1) were found based on ASD size, ductus arteriosus size, or oxygen saturation. The adverse effects of PGE(1) were apnea in 10 patients and fever in 19 patients. Neither seizures nor necrotizing enterocolitis was documented. Prostaglandin E(1) was successfully withdrawn for a minority of hemodynamically stable patients with d-TGA. No predictors for a successful weaning could be identified. Because apnea and fever are common side effects, withdrawal of PGE(1) after BAS may improve patient safety and comfort. In this patient group, if PGE(1) withdrawal was not well tolerated, it could be safely reinitiated. There were no serious side effects of PGE(1.).

Long-term subcutaneous morphine administration after surgery in newborns.

AIM: To analyze the management of newborns after major surgery receiving morphine subcutaneously and to identify possible side effects. METHODS: Morphine was administered via a subcutaneous catheter (Insuflon) in 20 newborns after major surgery. Side effects like hypotension, pain during morphine administration and local infection were noted. Morphine dose was adjusted according to the hospital guidelines with the Neonatal Infant Pain Score (NIPS) and the Finnegan withdrawal score. RESULTS: Surgery was performed at the median age of 38 5/7 weeks (range: 32 1/7-49 5/7 weeks). Before starting subcutaneous morphine administration, patients received intravenous morphine for a median of two weeks (range six days to seven weeks). All patients showed good pain relief with no severe side effects. Three patients reacted with crying to the first dose of subcutaneous morphine. No other side effects occurred. CONCLUSION: Subcutaneous application of morphine with the Insuflon catheter is an alternative to intravenous treatment of postoperative pain in neonates. In this small group pain relief was good and side effects were harmless.

Predictive factors for the success of noninvasive mask ventilation in infants and children with acute respiratory failure.

OBJECTIVE: Noninvasive mask ventilation (NIV) is a treatment option in acute respiratory failure in adults. This study was performed to determine prognostic variables for the success of NIV in a group of infants and children with respiratory failure for a wide range of reasons. DESIGN: Prospective, clinical study. SETTING: Multidisciplinary, neonatal-pediatric intensive care unit of a university teaching hospital. METHODS: Descriptive study of infants and children 80% after 1 hr of NIV predicted nonresponse with a sensitivity of 56%, specificity of 83%, and positive and negative predictive value of 71%. CONCLUSION: NIV can be successfully applied to infants and children with acute respiratory failure in the setting of a pediatric intensive care unit. The level of Fio2 after 1 hr of NIV may be a predictive factor for the treatment success.

Arterial thromboembolic complications in critically ill children.

PURPOSE: To evaluate incidence and characteristics of arterial thromboembolic complications in critically ill children. MATERIALS AND METHODS: Hospital records of all consecutive patients with arterial thromboembolic events (ATEs) occurring in the pediatric intensive care unit (PICU) from January 1997 to August 2001 were reviewed. Data collected included demographics and location, treatment modalities and outcome of ATEs. RESULTS: Fifty-four ATEs in 51 children (median age, 14 days) were identified, reflecting an incidence of 1.2% of all PICU patients. Arterial thromboembolic events were located in peripheral arteries in 52 (96%) cases and were associated with indwelling arterial catheters (n=26) or cardiac catheterization (n=26). The remaining 2 ATEs were located in the left ventricle and cerebral arteries, respectively. Therapy consisted of heparin (n=51), thrombolysis (n=22), oral anticoagulation (n=12), and aspirin (n=34). Complete resolution was noted in 33 (70%), partial resolution in 10 (21%), and no resolution in 4 (8.5%) cases. Bleeding complications occurred in 1 patient treated with heparin and in 12 (54%) of the 22 patients receiving thrombolytic therapy. CONCLUSIONS: Arterial thromboembolic events are frequent complications of PICU, particularly affecting neonates, and mostly associated with catheters. Studies to determine safe and effective prophylactic and treatment modalities of ATEs in children are required.

Evaluation of a new combined transcutaneous measurement of PCO2/pulse oximetry oxygen saturation ear sensor in newborn patients.

OBJECTIVE: Arterial oxygen saturation (Sao(2)) and arterial carbon dioxide partial pressure (Paco(2)) are 2 of the most important respiratory parameters in the treatment of critically ill neonates. Noninvasive monitoring of these parameters is desirable for continuous estimating of the respiratory status and reducing blood loss because of repeated blood gas analyses. Transcutaneous measurement of Pco(2) (Ptcco(2)) represents a simple and noninvasive technique for continuous monitoring of ventilation. However, sensor preparation, positioning, taping, and repeated changes of the sensor location make the handling difficult and complicate its use in the neonatal care unit. Recently, a new sensor for combined assessment of pulse oximetry oxygen saturation (Spo(2)) and Ptcco(2) has been introduced (TOSCA Monitor; Linde Medical Sensors, Basel, Switzerland). The monitor combines pulse oximetry and Ptcco(2) measurement in a single ear sensor, which works at 42 degrees C to enhance blood flow in capillaries below the sensor. METHODS: In a prospective, open, nonrandomized study of 60 ill neonates, the new ear sensor for combined assessment of Spo(2) and Ptcco(2) at 42 degrees C was tested. The sensor was adapted to the ear of a neonate with a Varihesive layer (Conva Tec; Princeton, NJ). Data obtained from the ear sensor were compared with Spo(2 Finger/Heel), Sao(2), and Paco(2) obtained from arterial blood gas in 30 patients and with a capillary blood gas in an additional 30 patients using Bland Altman bias analysis. Data are presented as median (range). RESULTS: The postconceptional age of the patients was 38.3 weeks (range: 28 5/7-40 5/7) in the arterial group and 37.9 weeks (range: 29 6/7-41 0/7) in the capillary group. Age of the newborns studied was 3.5 days (range: 1-28) in the arterial blood sample group (n = 30) and 6 days (range: 2-28) in the capillary blood sample group (n = 30). Patient weight was 3.02 kg (range: 1.5-4.5) in the arterial group and 2.76 kg (range: 1.0-3.71) in the other group. Three patients had weights of <1500 g. Twenty-one of 60 patients were conventionally ventilated, 4 patients received high-frequency oscillation, and 35 were not ventilated. Mean difference (bias) and precision (2 SD of the mean difference) between Ptcco(2 TOSCA) and Paco(2) were -0.44 kPa (-3.21 mm Hg) and 0.82 kPa (6.02 mm Hg) and between Ptcco(2 TOSCA) and Pcapco(2) were -0.09 kPa (-0.67 mm Hg) and 1.11 kPa (8.07 mm Hg), respectively. Spo(2) assessment by the TOSCA revealed slightly higher values compared with Sao(2) (bias: -0.48%), whereas Spo(2) (Finger/Heel) values were slightly lower than Sao(2) (bias: 0.52%). CONCLUSION: The TOSCA monitor with the ear sensor adapted to ears of neonates allows reliable estimation of Sao(2) and Paco(2). A potential benefit is the reduction in motion artifacts because of less head movement in newborns and that only a single cable leads form the patient to the monitor. In addition, the sensor is not removed for chest radiograph or for nursing the infant on his or her parent's lap. Long-term studies in a large population with continuous measurements are required to confirm these preliminary findings and to elucidate the benefits in detection of respiratory deterioration and the potential side effects of this sensor.

Efficacy and safety of intravenous amiodarone for incessant tachycardias in infants.

UNLABELLED: Amiodarone is an effective anti-arrhythmic agent for the treatment of supraventricular and ventricular tachycardias. The safety and efficacy of intravenous amiodarone has been described in adults and children but only to a limited extent in infants. The purpose of this study was to evaluate the safety and efficacy of intravenous amiodarone in infants. Between February 1994 and June 2001, 23 infants with a median age of 8 days (range 1-300 days) with life-threatening incessant tachycardias (17 supraventricular, 6 ventricular) were treated with intravenous amiodarone as single anti-arrhythmic agent. At presentation, 22 infants were haemodynamically unstable. Amiodarone was given as an intravenous loading dose of 5 mg/kg over 1 h followed by an intravenous maintenance dose of 5 micro g/kg per min with stepwise increase up to 25 micro g/kg per min until arrhythmia control or side-effects occurred. Amiodarone was effective in 19 infants, partially effective in three and ineffective in one infant. The median time until arrhythmia control was 24 h (range 1-96 h) and the median maintenance dosage 15 micro g/kg per min (range 5-26 micro g/kg per min). Electrophysiological side-effects necessitating dose reduction comprised of sinus bradycardia in two patients. Hypotension in one patient resolved after dose diminution. Neurological side-effects consisted of choreatic movements in one infant, which resolved over time. Amiodarone administration was stopped in one patient with elevated liver enzymes. CONCLUSION: intravenous amiodarone is a safe and effective therapy for life-threatening incessant tachycardias in infants.

Flush volumes delivered from pressurized bag pump flush systems in neonates and small children.

BACKGROUND: The aim of this study was to measure the volumes of fluid delivered with a fast flush bolus from a flow regulating device. METHODS: In-vitro fast flush bolus volumes, the volumes delivered from a bag pump flush system while opening the flow regulating device for 1, 2 or 5 s, were gravimetrically measured through a 22-G and a 24-G cannula. In-vivo 1- and 2-s fast flush bolus volumes and the volume required to purge the tubing between stopcock and arterial cannula from visible blood after blood sampling were recorded in 12 anaesthetized neonates and infants (mean age 2.17 +/- 1.97 months, range 0.26-5.37 months) with a 24-G radial arterial cannula by continuously weighing the bag pump flush system at manometer pressures of 100, 200 and 300 mmHg. RESULTS: In-vitro fast flush bolus volumes ranged from 0.23 +/- 0.04 ml (1-s, 100 mmHg, 24-G cannula) to 2.95 +/- 0.38 ml (5-s, 300 mmHg, 22-G cannula). Volumes were larger using a 22-G cannula than a 24-G cannula (P < 0.01) and increased with longer flushing periods (P < 0.0001) and higher manometer pressures (P < 0.0001). In-vivo 1- and 2-s fast flush bolus volumes correlated well with driving pressures (infusion pressure minus mean arterial pressure) (r2 = 0.81/0.72). 1-s fast flush bolus volumes delivered (ml) were 0.0025 x mmHg driving pressure and 2-s fast flush bolus volumes delivered (ml) were 0.0043 x mmHg driving pressure. The mean volume delivered to purge blood from the arterial pressure tubing was 0.94 +/- 0.18 ml (range 0.61-1.34 ml). CONCLUSIONS: Fast bolus flushing from pressurized infusion bag systems, using the flow regulating device tested, can be applied during neonatal and paediatric anaesthesia without delivering uncontrolled amounts of fluid.

Does critical incident reporting contribute to medication error prevention?

UNLABELLED: Medication-related critical incidents (CIs) comprise harmful and potentially harmful events. The aim of CI monitoring is quality improvement through system changes. In a prospective survey, we analysed our drug-related CIs of the year 2001 with an emphasis on how they contributed to system changes. A voluntary, anonymous, non-punitive CI reporting was used. The study was performed in a multidisciplinary, 23-bed, neonatal-paediatric intensive care unit (ICU). CI severity was graded: minor (no interventions required), moderate (requiring routine therapy, available outside the ICU), major (need for therapeutic interventions specific to the ICU). There were 284 drug-related CIs, 76% (95% confidence interval 71%-81%) of minor, 19% of moderate and 5% of major severity. A total of 24 CIs were potentially life threatening (if not detected). Some 27% of CIs were intercepted, 17% before preparation and 10% before administration of the drug to the patient. There was a negative correlation between median delay (from CI to detection) and mean severity of the different drug classes involved (P = 0.027). As to the impact on quality, 46 CIs were followed by system changes and 63% (95% confidence interval 49%-77%) of these CIs were of minor severity. Examples of system changes are: double checking for potentially harmful drugs, standardised prescription form and contact to the national drug control agency regarding misleading drug labels. CONCLUSION: most of the system changes were based on minor critical incidents which were often detected only after a longer period of time. This shows the value of our "low-threshold" critical incident monitoring. Repeated checks along the drug delivery process (prescription, preparation, administration) are an important means to reduce adverse drug events.