A cost-analysis of suspected adverse drug reactions in a hospital emergency ward.

The main objective of this study was to analyse the minimum direct cost to the Public Health System (PHS) of diagnosing and treating those patients attended to in the emergency ward (EW) for suspected adverse drug reaction (ADR). The cases were collected during March 1995 in the emergency ward of a 900-bed tertiary teaching hospital that covers 900,000 inhabitants. ADR was considered according to the WHO definition. The following EW costs were used: EW physician visit 78.5 ecus (1 ecu=156 pesetas), thorax or abdomen radiograph 21.5 ecus, computerized tomography 112.7 ecus, endoscopy 48 ecus, specialist physician visit 62 ecus. Three types of laboratory costs were considered: block of biochemical tests 16 ecus, biochemical tests with blood count 22.5 ecus, and biochemical tests with blood count and coagulation study 41.6 ecus. Pharmacotherapy of the ADR and changes in patient's usual drug therapy due to ADR were estimated. For patients admitted in the hospital, a per day cost of 391 ecus was considered. A mean ADR cost per organ or system affected (cutaneous, metabolic, gastrointestinal, nervous) was computed. The main conclusion of this study is that ADR, apart from inflicting damage on the patients, also incur PHS costs. The quantity of 42,732 ecus during the month of March, in a single hospital, can seem small when compared with the cost of some diseases such as AIDS or ischemic heart disease. But remembering that about 40% of all ADR attended to in hospitals is avoidable, a decrease of 40% could produce an annual saving of 205,216 ecus to the hospital, which is twice the annual budget of the Pharmacovigilance Centre of the Basque Country. Pharmacovigilance centres should include cost analysis of ADR among their objectives, to provide health systems managers with enough information to implement those measures that will result in a better utilization of the scarce resources of the Public Health System.

[Adverse drug reaction surveillance in pediatric and adult patients in an emergency room]. / Monitorización en un servicio de urgencias de reacciones adversas causadas por medicamentos en niños y adultos.

BACKGROUND: Adverse drug reactions (ADR) in the pediatric emergency room of a tertiary care hospital in Spain are described and compared with the adult ward. METHODS: Identification of cases was carried out through review of admission diagnoses and selection of those included in a previous list of diagnoses considered as possible ADR, that were thereafter verified. RESULTS: During 25 months, in 47.107 pediatric consultations were detected 451 cases as suspicious of ADR (0.96%). The ADR was moderate in 29 and serious in 1, being hospitalized 4. In adults, there were 68,431 consultations, and 704 cases detected (1.03%); moderate 218, serious 34 and mortal 1, being hospitalized 101. The most common reactions were dermatological (43.9% in children, 19.5% in adults) and of digestive nature (28.5 and 36.6%, respectively). The drugs most frequently involved in children were antimicrobials (49.5%), drugs used in respiratory illnesses (19.9%), non-steroid anti-inflammatory drugs (NSAID) (10.4%) and vaccines (9.2%); only two recently marketed products were involved. In adults, drugs involved were NSAID (28.2%), cardiovascular drugs (15.9%), antimicrobials (14.5%) and drugs active in digestive system (11.1%). There were 10 cases of hypoglycemia in diabetic adults, probably by interaction of hypoglycemic agents with angiotensin-converting enzyme inhibitors, and 7 cases of gastrointestinal hemorrhage associated with ketorolac, that generated an alert; 12 recently marketed products were involved. CONCLUSIONS: Intensive monitoring in emergency ward measures ADR problem, estimates underreporting, but it has a moderate value to generate alert or to survey new products.

Characteristics of serum protein binding of felodipine.

The protein binding of felodipine in concentrations ranging from 3 to 65 nmol L-1 has been characterized in pooled serum, and in isolated human plasma proteins: albumin (HSA) and a1-acid glycoprotein (AAG). Protein binding was determined by an ultrafiltration technique using an Amicon Micropartition System. Serum protein binding of felodipine (16 nmol L-1) was measured in five groups of individuals: I: healthy subjects (n = 16). II: patients with chronic renal disease before and after dialysis (n = 10). III: patients with liver disease (n = 9). IV: diabetics Type I (n = 10) and Type II (n = 12) and V: cancer patients (n = 12). Concentrations of HSA, AAG, lipoproteins and non esterified fatty acids (NEFA) were also measured. The drug was extensively bound in pooled serum and the protein binding was essentially unchanged over the concentrations of felodipine studied (99.60 +/- 0.31% at 3 nmol L-1; 99.70 +/- 0.15% at 65 nmol L-1). In albumin solution (40 g L-1) felodipine was also highly bound. The mean of percentage bound was not significantly different from that in serum and was also independent of the felodipine concentrations (98.57 +/- 0.35 at 3 nmol L-1; 98.31 +/- 0.90 at 65 nmol L-1). The extent of binding to AAG was significantly lower than in serum (p < 0.01) and HSA (p < 0.01) and was independent of felodipine concentrations (85.64 +/- 2.25 at 3 nmol L-1; 85.68 +/- 2.3 at 65 nmol L-1). The percentage of bound felodipine in group II (before dialysis) was significantly lower than in group I (p < 0.001). The variability in the percentage of bound felodipine was greater in group II, before dialysis, than in the rest of the groups. After dialysis, protein binding was similar to that in group I. HSA did not change and AAG was increased. NEFA was significantly higher after dialysis when compared with group I (p < 0.01). In vitro carbamylation of serum did not change felodipine protein binding. HSA was decreased significantly in group III patients (p < 0.05). However, protein binding did not change. Binding of felodipine in the rest of the groups was not significantly different from that in group I. Linear regression analysis of the data for all individuals indicated that the binding of felodipine was related to serum lipoproteins and that age, HSA, AAG, and NEFA were not significant determinants of binding.

The effect of changes in gastric pH induced by omeprazole on the absorption and respiratory depression of methadone.

The effect of omeprazole (2 mg kg-1 i.v.) on respiratory depression induced in rats by acute oral methadone administration (5 mg kg-1) was examined and compared with control animals that only received methadone. Quantitative assessments of arterial Pco2,Po2, pH, and respiratory rate were employed as criteria for evaluation. Intragastric pH was measured in each rat immediately before and 2 h after methadone. Plasma concentration of methadone was measured for 3 h. The relationship between drug effect and the systemic bioavailability of methadone, measured as the area under the plasma concentration-time curve (AUC0-180), was also evaluated. The intensity of the methadone-induced respiratory depression was significantly greater in the omeprazole group than in control rats. A significant variation (p < 0.01) in all respiratory parameters was detected from 30 to 120 min after methadone. Omeprazole caused a significant increase in methadone levels (Cmax = 156 +/- 6.5 ng mL-1 against 51 +/- 5.8 ng mL-1 in control; p < 0.05). AUC0-180 was higher (p < 0.05) after omeprazole treatment (18.6 +/- 1.4 micrograms mL-1 min) than in control (6.8 +/- 0.6 microgram mL-1 min). Two hours after treatment with omeprazole, intragastric pH values were significantly elevated (4.7 +/- 0.1 against 2.2 +/- 0.04) and continued increasing, being 6.4 +/- 0.1 at the end of the experiment. Correlation was observed between intragastric pH and the area under the effect- (respiratory depression-) time curve (r = 0.74; p < 0.001). A relationship between plasma methadone levels at 120 min and gastric pH (r = 0.92; p < 0.001) was detected. A significant correlation between the area under the effect-time curve (0-120 min) and AUC0-180 has been also observed (r = 0.90; p < 0.01). These pharmacokinetic and pharmacodynamic changes could be gastric pH dependent because they were mimicked when gastric pH was experimentally modified by bicarbonate whereas opposite results were obtained with acidic pH2 solution.

Pharmacokinetics and pharmacodynamics of penbutolol in healthy and cancer subjects: role of altered protein binding.

The pharmacokinetic and pharmacodynamic profiles of penbutolol were examined in healthy volunteers and in cancer patients using a pharmacokinetic/pharmacodynamic (pk/pd) model. After receiving a 40 mg single oral dose of penbutolol, the absorption rate constant, apparent volume of distribution and serum clearance of penbutolol were found to be reduced in the cancer group. Changes in the disposition of the conjugate metabolite were also observed in the cancer patients. Penbutolol unbound fraction in serum was statistically decreased (p < 0.005) in the cancer group, according to the increase in the serum levels of alpha 1-acid glycoprotein seen in that group (p < 0.05). The pharmacodynamic effect of penbutolol was measured as the reduction in heart rate (HR); in healthy volunteers, a linear relationship (p < 0.01) between effect and penbutolol serum concentrations (total or unbound) was found. In contrast, in cancer patients, values of HR did not vary statistically in respect to baseline values. These results show that in cancer patients, a change in the pharmacokinetics of penbutolol occurs (associated with changes in drug protein binding), together with an alteration in the pharmacodynamics.

Resistance to atracurium in rats with experimental inflammation: role of protein binding.

The influence of altered protein binding on the neuromuscular effect of atracurium has been studied in rats with experimental inflammation induced by subcutaneous injection of turpentine oil. Doses of atracurium ranging from 0.45 to 1.5 mg.kg-1 were administered to control (n = 30) and to experimental inflammation induced rats (n = 30). Neuromuscular transmission was monitored by recording the twitch tension of the tibialis-anterior muscle elicited by stimulation of the sciatic nerve. Three effect parameters were recorded: (i) intensity of the effect, measured as percentage depression of baseline twitch tension, (ii) duration of drug action (min) and (iii) recovery time (min). The dose-intensity of the effect relationship was modelled using a sigmoid Emax model. The ED50 (effective dose eliciting 50% of the maximum effect) was significantly increased (P < 0.01) in the inflammation group as compared to the control group (0.94 vs. 0.68 mg.kg-1). This change was reflected in a shift of the dose-response curve to the right in the pretreated rats. For equipotent doses ED95 (defined as the effective dose eliciting 95% of maximum effect), no differences were found in recovery time and duration of action between the two groups of rats. Mucoproteins levels (index of alpha 1-acid glycoprotein (AAG) and protein binding were significantly increased in rats with experimental inflammation as compared to control rats. Based on these results, altered serum protein binding of atracurium appears to be responsible, at least in part, for the resistance to atracurium.

Changes in the analgesic effects of mianserin associated with altered plasma protein binding in experimental cancer.

In a group of mice bearing experimentally induced tumors, the protein binding of mianserin in vitro was measured and compared with a control group. The analgesic effect and the brain uptake of drug was also compared with a control group after an intraperitoneal dose of mianserin. The unbound percentage of mianserin in the plasma of mice with experimental cancer decreased with respect to control animals (5.20 +/- 0.12 vs 6.06 +/- 0.26; p<0.05) and alpha1-acid glycoprotein (AAG) levels, measured as plasma mucoprotein concentrations, were significantly increased (p<0.05). The brain/plasma drug concentration ratio of mianserin decreased in mice with experimental cancer when compared with control mice (1.11 +/- 0.03 vs 1.42 +/- 0.10; p<0.02). In both groups of mice, the mianserin analgesic effect was evaluated by the hot plate test after intraperitoneal drug administration. When the analgesia response-dose curve (0-60 mg/kg) was studied, a significant decrease in the response in mice with experimental cancer versus control mice was observed. These results suggest that resistance to the mianserin analgesic response may occur in animals with cancer disease. This resistance may be associated, in part, with an altered plasma protein binding, but other mechanisms could be involved.

Differential effect of cancer on the serum protein binding to mianserin and imipramine.

Protein binding of mianserin and imipramine in vitro was determined in sera from 10 patients with cancer and from 28 drug-free normal subjects. alpha 1-acid glycoprotein (AAG) concentrations ranged from 0.91 +/- 0.04 g/l in control subjects to 2.17 +/- 0.18 g/l in cancer patients. Albumin concentrations ranged from 55.80 +/- 1.68 g/l in control subjects to 39.71 +/- 4.40 g/l, respectively. Serum samples containing concentrations of 100 ng/ml for mianserin and 500 ng/ml for imipramine were ultrafiltered and the free concentration were measured with scintillation spectrophotometer. The mean free percentage of mianserin was significantly less in patients with cancer (8.70 +/- 0.29% in patients vs 14.30 +/- 0.50% in control subjects P < 0.001). A multiple regression analysis revealed a significant contribution of plasma AAG (r2 = 0.56, P < 0.01), but not of albumin to the overall variability in mianserin binding. No correlation was observed between protein binding of imipramine and AAG concentrations in serum of cancer patients. No significant changes were observed for protein binding of imipramine in cancer patients as compared with control subjects. Our results suggest that for antidepressant (AD) drugs, of which the binding depends on AAG, variability in protein binding could be expected in cancer patients. Thus, in cancer therapy, changes in analgesic doses could be necessary with this kind of antidepressant drug.

Influence of vitamin C on the absorption and first pass metabolism of propranolol.

The effect of ascorbic acid on the availability of propranolol has been examined. After oral administration of propranolol 80 mg with or without ascorbic acid pretreatment (2 g), the plasma concentrations and urinary excretion of propranolol and its metabolites, 4-hydroxy-propranolol and propranolol-conjugated, were determined by HPLC. Compared to controls, vitamin C decreased the maximum concentration of propranolol from 463 to 334 nmol.l-1, and the area under the propranolol concentration-time curve (from 0 to 24 hours) from 3.13 to 1.96 mumol.l-1.h. The time to reach maximum propranolol concentration was increased from 1.9 to 2.7. The total amount of drug recovered in urine has also significantly diminished (from 12.6 to 4.29 mg). No change in elimination rate was observed, indicating that ascorbic acid had affected both the absorption process and the first pass metabolism. The heart-rate decreased less when propranolol was administered with ascorbic acid in comparison to control subjects, although this interaction has little biological importance.

The extent of sodium and water administration associated with i.v. drug infusions in a surgical unit.

OBJECTIVE: To quantify sodium and water administered to surgical patients but not actually prescribed by the physician. METHODS: Computerized medication records of 1208 surgical patients who were operated on in 1992 were analysed retrospectively. Only prescriptions for intravenous (i.v.) drugs and fluids were selected. RESULTS: Thirteen of 143 i.v. drugs made up 68.3% of prescriptions for i.v. drugs. Patients received a median of three i.v. drugs daily. Patients were on i.v. therapy for a mean of 4.6 +/- 6.0 days (range 1-140, median 3 days). On 24% of patient-days, patients received more than 100 mEq of sodium which was not consciously prescribed, and on 20.4% of patient-days they had an extra daily intake of 750 ml of water or more. One fourth of the total sodium and more than 15% of fluid administered to patients was in the form of NaCl 0.9%, used as a vehicle for i.v. drugs. CONCLUSIONS: There is a considerable amount of sodium and water administered to patients without a specific prescription. The promotion of knowledge about the correct administration of i.v. drugs is an important task for the clinical pharmacist. If this is achieved, it will allow for an optimal administration of drugs as well as the prevention of possible side effects from an intended sodium and water administration.

[Characterization of propofol binding to plasma proteins and possible interactions]. / Caracterización de la fijación de propofol a las proteínas plasmáticas y posibles interacciones.

OBJECTIVES: a) To study the binding of propofol to proteins in plasma samples from healthy volunteers and in solutions of albumin and alpha 1-acid glycoprotein (AGA); b) to describe the nature of the bond and possible interactions with other substances that are potential displacers: salicylate, phenylbutazone, sulfisoxazole, tolbutamide, sodium valproate, sodium oleate and penbutolol; c) to assess the effect of propofol on the binding of specific markers and possible binding sites in the following proteins: 14C-warfarin, 3H-diazepam, 3H-midazolam, 3H-imidazole, 3H-penbutolol and 3H-morphine. MATERIAL AND METHODS: The free fraction was obtained in all samples by ultrafiltration and measurement of the free concentration of propofol by liquid chromatography and of the markers by scintillation spectrometry. RESULTS: The free fraction of propofol in plasma was 0.98 +/- 0.12% and binding was not saturable. Albumin seems to play an important role (95% bound), whereas the participation of AGA was low (54% bound). Propofol did not affect the binding of any of the markers studied. Nor did the presence of other drugs at therapeutic plasma concentrations affect the binding of propofol. CONCLUSIONS: The binding of propofol to plasma proteins seems unlikely to cause drug interactions in clinical practice.

Intravesical BCG therapy of superficial bladder cancer: study of adverse effects.

Immunotherapy with intravesical bacillus Calmette-Guérin (BCG) has proved to be more effective than most chemotherapeutic agents in the prophylaxis and treatment of superficial bladder tumours. Side-effects, both local and systemic, are the main limitations against its use. With the aim of lowering the incidence and severity of side-effects, we started to use two vials of BCG, Connaught type, per instillation, instead of three vials, as recommended by the manufacturer. We prospectively reviewed adverse effects of BCG treatment at the lower dosage in 92 patients. Compared with other series, we found a similar incidence of adverse effects except for some local effects as haematuria which showed a higher incidence, but we also found a lower rate of tumour relapse. Four primary tumours were recorded during the study period. In our open study, a lower BCG intravesical dosage is not followed by a reduction in side-effects.

Decrease in penbutolol protein binding as a consequence of treatment with some alkylating agents.

The effect of in vitro treatment of serum with the alkylating agents carmustine (BCNU) and mechlorethamine on the protein binding of penbutolol, a basic agent mainly bound to alpha 1-acid glycoprotein (AAG), was investigated. The free fraction of penbutolol increased significantly (P < 0.001) after the treatment of serum with BCNU (5.27+ +/- 0.47%) and with mechlorethamine (5.23% +/- 0.17%), being 1.98% +/- 0.18% in serum not treated with BCNU or mechlorethamine. In addition, after incubation with BCNU (2 h), the free fraction of penbutolol continued increasing (10.96% +/- 0.70% vs 5.27% +/- 0.47% at time 0; P < 0.001), whereas it remained unchanged after incubation with mechlorethamine. Moreover, dialysis against saline for 24 h did not restore the free fraction of penbutolol, which increased after treatment with carmustine (9.05% +/- 1.24% vs. 11.04% +/- 1.55%, nondialyzed). We concluded that the treatment of cancer patients with alkylating agents could alter the serum proteins and modify their binding capacity, and this should be taken into account in the simultaneous treatment of these patients with other basic drugs like penbutolol, e.g., methadone.

Netilmicin kinetics in urology.

Plasma kinetics of netilmicin, an aminoglycoside antibiotic was studied in 62 patients undergoing urologic surgery. Despite the use of a standard 100 mg-dose, no toxic levels were achieved except in one patient. A poor correlation was found between netilmicin plasma elimination constant and creatinine clearance (r = 0.34, p = NS). We can conclude that the prediction of netilmicin plasma concentrations is not possible using only demographic patient's data. The monitoring of netilmicin levels should be performed in long-term treatments but not in 4-dose regimes such as in urologic prophylaxis.

Unchanged protein binding of penbutolol in renal insufficiency: a possible role of carbamylation.

The effect of in vitro carbamylation of serum protein with potassium cyanate on protein binding of penbutolol, a basic agent exclusively bound to alpha 1 acid glycoprotein (AAG), was investigated. Carbamylation of serum resulted in a weak increase on free fraction of penbutolol (4.45 +/- 0.54% before carbamylation vs 5.66 +/- 0.40% after; p < 0.025). Parallelly, potassium cyanate added to pure AAG and incubated for 90 min induced carbamylation of this protein (38 mumoles of 14C cyanate incorporated per gram of protein). A study in serum from patients with chronic renal disease (pre and postdialysis) showed no changes in protein binding of penbutolol, although AAG levels were significantly higher. However, Scatchard [1949] plot for penbutolol binding to serum from renal patients (both pre and postdialysis) showed a decrease in affinity constant (nKa = 11.13 x 10(5) M-1 in healthy volunteers, vs 5.56 x 10(5) M-1 in patients before dialysis and 4.57 x 10(5) M-1 after dialysis). We concluded that carbamylation of serum AAG in uremic patients could explain, in part, the absence of changes in protein binding of any basic drugs in this pathological condition. It appears that a decreased affinity constant could balance the effect of increased AAG levels.

The influence of renal failure on the kinetics of intravenous midazolam: an "in vitro" and "in vivo" study.

The effect of renal failure upon the "in vitro" binding of midazolam, a new water-soluble short-acting benzodiazepine, has been studied in man. An increase of its free fraction (ranging from 2.52 to 5.17%) in serum from uremic patients was observed. A similar situation was originated in rabbits by administering uranyl nitrate (2 mg/Kg i.v.) and posterior hypnosis with midazolam. Uremic rabbits showed a marked increase in the free concentration of midazolam in serum (ranging from 8.9 to 13.7 micrograms/ml) and in midazolam brain levels (156.2 micrograms/g in cortex vs 84.5 micrograms/g in control animals). A positive correlation between brain and serum free concentration of midazolam was also observed. It is concluded that in renal patients more unbound drug is available to produce central nervous system effects, and a decrease in intravenous dose of midazolam could be recommended in this clinical situation.

Influence of changes in protein binding on the central activity of antidepressants.

The central effect (expressed as analgesic response), protein binding and brain uptake of mianserin were measured in mice receiving drug intraperitoneally. A significant decrease of the central effect of mianserin (30 mg kg-1) was seen in mice with experimental inflammation when compared with control animals (reaction time (s) = 12.12 +/- 1.22 vs 25.56 +/- 2.92; P less than 0.001) and the dose-analgesia response curve (10-60 mg kg-1) was significantly shifted to the right in mice with inflammation. In serum of mice with inflammation, unbound concentration of mianserin was decreased from 19.37 +/- 0.73 to 17.83 +/- 0.30% (P less than 0.05) and seromucoid levels were significantly increased (P less than 0.001). Following the intraperitoneal administration of 30 mg kg-1 of mianserin, brain uptake decreased in diseased mice when compared with control animals (P less than 0.02), suggesting that the decrease in analgesia was secondary to a decrease in drug delivery to the brain because of increased protein binding.