The relationship between extracorporeal circuit prime, albumin, and postoperative weight gain in children.

OBJECTIVES: This study evaluated postoperative weight gain in children who received albumin versus crystalloid prime for cardiopulmonary bypass (CPB). DESIGN: A retrospective case-controlled study. Children whose extracorporeal (EC) circuit prime contained albumin (group 1) were matched with those whose prime contained only crystalloid (group 2) on the basis of age, weight, and surgical repair. SETTING: A university-based medical center. PARTICIPANTS: Seventy-six children (newborn to 4 years of age) who underwent CPB for correction of a congenital heart anomaly from 1993 to 1995. Group 1 underwent surgery from October 1994 to September 1995, and group 2 from February 1993 to September 1994. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Group 1 had less weight gain on postoperative days (PODs) 1, 2 and 3 compared with group 2 (p = 0.04 on POD 1). Albumin (grams per milliliter) prime and prime volume in milliliters per kilogram were the best predictors of weight gain (p < 0.004), with prime volume being the more important. Children who weighed less than 7.5 kg received more prime volume and had greater weight gain than children who weighed 7.5 kg or greater on PODs 1, 2, and 3 (p < 0.02). CONCLUSION: Data suggest that adding albumin to the EC circuit prime and minimizing the prime volume will result in less postoperative weight gain. Further prospective study with a larger sample is warranted to determine whether albumin prime offers other clinical benefits.

The FLACC: a behavioral scale for scoring postoperative pain in young children.

PURPOSE: To evaluate the reliability and validity of the FLACC Pain Assessment Tool which incorporates five categories of pain behaviors: facial expression; leg movement; activity; cry; and consolability. METHOD: Eighty-nine children aged 2 months to 7 years, (3.0 +/- 2.0 yrs.) who had undergone a variety of surgical procedures, were observed in the Post Anesthesia Care Unit (PACU). The study consisted of: 1) measuring interrater reliability; 2) testing validity by measuring changes in FLACC scores in response to administration of analgesics; and 3) comparing FLACC scores to other pain ratings. FINDINGS: The FLACC tool was found to have high interrater reliability. Preliminary evidence of validity was provided by the significant decrease in FLACC scores related to administration of analgesics. Validity was also supported by the correlation with scores assigned by the Objective Pain Scale (OPS) and nurses' global ratings of pain. CONCLUSIONS: The FLACC provides a simple framework for quantifying pain behaviors in children who may not be able to verbalize the presence or severity of pain. Our preliminary data indicates the FLACC pain assessment tool is valid and reliable.

The efficacy and cost of aprotinin in children undergoing reoperative open heart surgery.

We performed a prospective, randomized, placebo-controlled, double-blind trial to assess the efficacy of aprotinin in 61 children (median age 3.7 yr) undergoing reoperative open heart surgery (OHS). Three demographically similar groups were studied: large-dose aprotinin (ALD), small-dose aprotinin (ASD), and placebo (P). Over the first 24 postoperative hours fewer patients in the aprotinin groups received packed red cells (ALD, 53%; ASD, 89%; and P, 95%; P = 0.001), platelets (ALD, 32%; ASD, 50%; and P, 65%; P = 0.04), and fresh frozen plasma (ALD, 16%; ASD, 17%; and P, 60%; P = 0.003) than placebo patients. Most importantly, aprotinin patients had fewer exposures to banked blood components (ALD, median 1 U; and ASD, median 2 U) than P (median 6 U; P = 0.001), with no difference in overall complication rate. Use of aprotinin was associated with a savings in the patient charges for blood components, operating room time, and duration of hospitalization. In conclusion, aprotinin decreased the number of units of banked blood components used during the first 24 postoperative hours in reoperative pediatric OHS. Aprotinin thus decreases the risks associated with exposure to banked blood components and reduces hospital charges.

Age-related differences in heparin sensitivity and heparin-protamine interactions in cardiac surgery patients.

OBJECTIVE: The present study was conducted to determine how children and adults differ (it at all) with respect to sensitivity to heparin activity and heparin-protamine interactions during cardiac surgery requiring cardiopulmonary bypass (CPB). DESIGN: A prospective study of both children and adults undergoing CPB. SETTING: A tertiary care academic medical center between July 1992 and October 1994. PARTICIPANTS: Ninety patients who had cardiac or aortic arch surgery using CPB. The median age of the entire study sample was 15.8 years (range 2 months to 72 years). INTERVENTION: Data were obtained using the Medtronic Hemotec Hepcon Hemostasis Management System (Englewood, CO). An ex vivo heparin dose-response (HDR) curve was generated for each patient before skin incision to determine the target heparin concentration (THC) needed to achieve an activated coagulation time (ACT) of at least 480 seconds. Protamine dose was determined on the basis of whole blood heparin concentration estimated by means of a heparin-protamine titration. MEASUREMENTS AND MAIN RESULTS: The study population was divided into four groups based on age: infants (< 1 year), preschool (1 to 5 years), school-age (5 to 14 years) and adults (> 14 years). The mean +/- SD THC for the preschool group was 4.0 +/- 1.1; for infants, 3.3 +/- 0.7; for school-age, 3.1 +/- 0.7; and for adults, 3.4 +/- 0.7. The initial dose of heparin needed to achieve this THC (mean +/- SD) was significantly higher in infants (578 +/- 220 U/kg) and preschool children (477 +/- 159 U/kg) than in school-age children (327 +/- 57 U/kg) and adults (332 +/- 64 U/kg). The ratio of protamine to heparin was significantly higher in adults (1.4 +/- 0.5) and school-age children (1.3 +/- 0.6) than in infants (1.1 +/- 0.7) and preschool children (1.1 +/- 0.4). CONCLUSIONS: Pre-school children are less sensitive to heparin but also display a wider range of sensitivity. The data in this study support the use of 300 U/kg of heparin before CPB in patients > or = 5 years but suggest that heparin requirements may be greater in the younger patient who may require as much as 500 U/kg to achieve what is believed to be an appropriate target heparin concentration for initiating CPB.

The activated coagulation time: suitability for monitoring heparin effect and neutralization during pediatric cardiac surgery.

OBJECTIVES: To determine how the stage of surgery affects the relationship between activated clotting time (ACT) and heparin effect in children undergoing cardiac surgery using cardiopulmonary bypass (CPB) and to compare the results of ACT determinations made with two different coagulation timers using different clot detection technologies and activator compositions. DESIGN: Prospective, paired observation. SETTING: Tertiary care children's hospital affiliated with an academic medical center. PARTICIPANTS: Fifty-eight children scheduled for nonprimary cardiac surgery. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: ACTs were measured by two different commercially available automated coagulation timers (Hepcon Hemostasis Management System [HP] and the Hemochron model 801 [HM]) at four different time points over the course of cardiac surgery requiring CPB in patients ranging in age from 0.16 to 19 years. Simultaneous determinations of whole blood heparin concentration using heparin-protamine titrations were made as well. When the two methods of ACT determination were compared, baseline ACTs were not significantly different. HP ACT prolongation after heparin administration but before bypass was significantly less than HM ACT prolongation (median ACT range HM > or = 999 seconds; HP, 560 to 679; p = 0.006). Twenty-one percent of the HP ACTs and none of the HM ACTs fell below the 480 seconds required at this institution for the initiation of CPB (p = 0.008). Both instruments showed a significant further prolongation of ACT after initiation of bypass (median ACT range HP > or = 999 seconds; HM > or = 999; p < 0.001 for both), whereas the heparin concentration decreased significantly (before, 3.5 +/- 0.2 U/mL; after, 2.7 +/- 0.1; p < 0.001). After termination of CPB and heparin neutralization, no significant difference between the ACTs was found. However, four HP ACTs were > or = 999 seconds despite simultaneous HM baseline values and whole blood heparin concentrations of zero. Heparin concentration correlated with ACT prolongation using both the HM (Spearman p = 0.36; p = 0.02) and the HP (Spearman p = 0.57; p = 0.0025) instruments before, but not 10 minutes after, initiation of bypass. CONCLUSIONS: In pediatric cardiac surgery, the relationship between ACT and heparin concentration changes depending on when during the surgery the ACT is measured. ACT prolongation in children anticoagulated for CPB correlates poorly with heparin concentrations during CPB. HP and HM ACT tests are not interchangeable. The HM ACT is a better indicator of heparin neutralization than the HP ACT. On the other hand, continued prolongation of the HP ACT after heparin neutralization may be related to risk of postoperative hemorrhagic complications. If devices from different manufactures are freely substituted for each other, clinical practice may be altered in an uncontrolled manner.

Lumbar epidural morphine infusions for children undergoing cardiac surgery.

OBJECTIVE: To determine whether outcomes and costs in children undergoing cardiac surgery are affected by the method of postoperative pain management. DESIGN: Retrospective, case control. SETTING: Tertiary care children's hospital in a university setting. PARTICIPANTS: Two groups of children undergoing cardiac surgery for palliation or repair of congenital heart disease oer a 21-month period between January 1993 and September 1994. INTERVENTIONS: Lumbar epidural morphine infusions (LEM) in one group, and IV opioid (IVO) medication in the other for postoperative pain control. MEASUREMENTS AND MAIN RESULTS: Hospital courses of 27 LEM patients and 27 IVO patients were analyzed. In LEM patients, epidural catheters were placed following anesthetic induction, but before anticoagulation. A bolus of 50 micrograms/kg of preservative-free morphine sulfate was administered through the catheter, followed by a continuous infusion at 3 to 4 micrograms/kg/h for 22 to 102 (median, 46) hours postoperatively. The IVO patients received 50 micrograms/kg, IV, of fentanyl before incision followed by a continuous infusion at 0.3 microgram/kg/min. The fentanyl infusion rate was decreased to 0.1 microgram/kg/min postoperatively and maintained for 24 hours. Although the LEM group was demographically similar to the IVO group, times to tracheal extubation, transfer from the intensive care unit, and resumption of regular diet were significantly shorter in LEM patients. LEM and IVO patients received similar amounts of fentanyl during surgery (10.4 +/- 19.3 micrograms/kg/h v 13.7 +/- 8.1 micrograms/kg/h, p = 0.4). However, during postoperative recovery, LEM patients who were extubated late received significantly less supplemental opioid medication than IVO patients extubated late during the first 5 postoperative days. No complications related to dural puncture, bleeding into the epidural space, or respiratory depression were encountered. Pruritus and nausea/vomiting were the most commonly reported morbidities in both groups. Fifty-six percent (15/27) of LEM patients and 41% of IVO patients reported pruritus (p = 0.4). There was no significant difference in the incidence of nausea and vomiting between the groups (34% v 30%, respectively). CONCLUSIONS: Given the present methodologic limitations, the authors found improved outcomes only in LEM patients extubated late compared with IVO patients. Randomized, prospective studies to evaluate this conclusion and to determine the comparative efficacy and safety of LEM infusions are in progress.

Multiple organ dysfunction syndrome: end organ and systemic inflammatory response in a mouse model of nonseptic origin.

The authors measured the peripheral blood pro-inflammatory cytokine responses (tumor necrosis factor-alpha [TNF-alpha] and interleukin-6 [IL-6]) and related end organ responses ti intraperitoneal zymosan-saline suspension over 5 days in CD-1 mice. Other indicators of local and systemic inflammation included wet:dry weight ratios of lung, liver, kidneys, spleen, and bowel; peripheral blood hematocrit, white blood cell count, and platelet count; lung myeloperoxidase activity; lung protein leak; and bacterial translocation to liver, spleen, and mesenteric lymph nodes. The initial event in responses to zymosan A injection was a sharp rise in the peripheral blood TNF-alpha level, which crested within 1 h of injection. This response was followed by a peripheral blood leukocytosis also within 1 h, and a peak lung myeloperoxidase activity within 1-2 h of injection. The maximum lung permeability index occurred 8 h after injection (zymosan, .398 +/- .019 [n = 10]; saline vehicle, .266 +/- .007 [n = 10], p < .001) followed by the maximum lung wet:dry weight ratio, which occurred 18 h after injection. The peak wet:dry weight ratios for the other organs occurred between 12 and 24 h after injection as well. Peripheral blood IL-6 maxima followed TNF-a maxima after lags of several hours. The release of pre-formed TNF-a is likely the most proximal event following injection of zymosan, and may set in motion the processes that result in end-organ injury and secondary multiple organ dysfunction, particularly activation of leukocytes. The precise roles of TNF-alpha and IL-6 in the pathogenesis of end-organ injury, however, are not addressed.

Volatile anesthetic modulation of lung injury and outcome in a murine model of multiple organ dysfunction syndrome.

General anesthesia has been shown to have a significant impact on the inflammatory response. We hypothesized that lung pathophysiology will be attenuated in a mouse model of secondary multiple organ dysfunction syndrome (MODS) elicited by intraperitoneal zymosan suspension in saline. CD-1 mice were anesthetized for 6 h with either 1% halothane or 1.5% isoflurane in 30% oxygen in N2 carrier gas. Another group of mice was exposed to 30% oxygen in N2 carrier gas only. The inflammatory response to zymosan was quantified by measuring lung myeloperoxidase activity (neutrophil recruitment). Lung injury was estimated by determining the degree of lung permeability to radioactive albumin (permeability index). Unanesthetized injured mice exhibited maximal lung myeloperoxidase activity 2 h after zymosan injection (.671 +/- .07 delta OD.min-1), which was significantly attenuated (p < .01) in injured mice anesthetized with halothane (.369 +/- .054) and isoflurane (.324 +/- .055). The maximum lung permeability index occurred 8 h after injection in the unanesthetized, injured mice (.398 +/- .019), and was attenuated (p < .01) in injured mice anesthetized with halothane (.255 +/- .02) and isoflurane (.224 +/- .019). Histopathological findings corresponded to the quantitative myeloperoxidase and permeability index values. Halothane and isoflurane attenuate lung inflammation and injury in this mouse model of multiple organ dysfunction syndrome. This attenuation may be related to modulation of the inflammatory response by volatile anesthetics.

Halothane-oxidant interactions in the ex vivo perfused rabbit lung. Fluid conductance and eicosanoid production.

BACKGROUND: The present studies were undertaken to determine the interactions between halothane and oxidative injury with respect to endothelial integrity, as measured by pulmonary capillary filtration coefficient (Kfc), and production of arachidonic acid-derived mediators, in perfused rabbit lungs challenged with the oxidant tert-butyl-hydroperoxide (t-bu-OOH). METHODS: Isolated lungs were prepared from 27 New Zealand white rabbits (2-3 kg) and were perfused with Ca(2+)-free Krebs-Henseleit buffer solution. In group A (n = 9), lungs were ventilated with halothane 2.5% in carrier gas (5% CO2 in air); in group B (n = 9), with carrier gas alone; and in group C (n = 9), with carrier gas, but without injury. The lungs in the two injury groups (A and B) received four infusions of t-bu-OOH, 200 microM, over 1 min, directly into the pulmonary artery. The uninjured lungs received four infusions of vehicle (normal saline). Kfc was determined after each t-bu-OOH infusion. Concentrations of thromboxane B2 (TxB2) and 6-keto-prostaglandin F1 alpha were measured in samples of effluent perfusate obtained before and 30 s after the end of each infusion of t-bu-OOH. The wet/dry weight ratio of each pair of lungs was determined at the end of each experiment. RESULTS: Kfc progressively increased after each infusion of oxidant in group A when compared with the other two groups. Lung wet/dry ratios were elevated in group A (14.3 +/- 0.7) and group B (13.2 +/- 0.2) compared with group C (12.1 +/- 1.1). TxB2 production in group A (2206 +/- 263 pg.min-1.g-1 dry lung tissue) was greater than in group B (1413 +/- 127) by the final infusion of t-bu-OOH. CONCLUSIONS: Ex vivo perfused rabbit lungs ventilated with halothane exhibited, simultaneously, evidence of greater fluid conductance across the pulmonary capillary bed and production of thromboxane A2 when challenged with oxidant than did lungs ventilated with carrier gas. Both of these effects may be mediated by halothane-related enhancement of intracellular endothelial Ca2+ mobilization stimulated by intrapulmonary infusion of oxidant.

Halothane and isoflurane increase pulmonary artery endothelial cell sensitivity to oxidant-mediated injury.

Volatile anesthetics inhibit phagocytic cell function, yet little is known about their effects on target tissues or on the target tissue response to stimulated phagocytes. Experiments were performed to determine how exposure to halothane and isoflurane changes rat pulmonary artery endothelial cell (RPAEC) viability in response to the toxic oxygen metabolites produced by stimulated phagocytic cells. RPAECs were grown in monolayer culture. The monolayers were treated with phorbol myristate acetate (PMA) -stimulated human neutrophils at an effector-to-target ratio of 20:1 after equilibration with 0.4% or 1.7% halothane or 0.7% or 2.8% isoflurane. As measured by percent-specific release of incorporated 51Cr label (mean +/- SE), cytotoxicity in the presence of 1.7% halothane (75.3 +/- 3.4%) was significantly greater (P less than 0.02) than cytotoxicity in 5% CO2 in air (44.7 +/- 3.3%) and in 0.4% halothane (57.3 +/- 4.7%). Also, cytotoxicity in 1.7% halothane was significantly greater than in 0.4% halothane (P less than 0.02). The authors found that RPAECs incubated in isoflurane exhibited significantly greater release of 51Cr than cells incubated in the MAC equivalent concentrations of halothane: 78.2 +/- 2.6% in 0.7% isoflurane (P = 0.0004) and 83.8 +/- 1% in 2.8% isoflurane (P = 0.005). Because early neutrophil cytotoxicity has been found to be mediated primarily by hydroxyl radical (HO.) and hydrogen peroxide (H2O2), the authors measured H2O2 production by similar numbers of PMA-stimulated neutrophils under similar exposure conditions. In carrier gas, PMA-stimulated neutrophils produced 20.5 +/- 1.3 nmol H2O2.10(6) cells-1.h-1. At the higher concentrations of halothane, H2O2 production actually was inhibited in comparison with carrier gas (15.4 +/- 1.4 nmol H2O2.10(6) cells-1.h-1 in 1.7% halothane and 16.8 +/- 0.8 in 2.8% halothane), but the degree of inhibition did not reach statistical significance. In isoflurane, however, H2O2 production was not different from that seen in carrier gas. In other experiments, the monolayers were treated with 0, 200, 500, and 1,000 microM H2O2 after equilibration with 0.4%, 1.7%, and 2.8% halothane or 0.7%, 2.8%, and 5% isoflurane in 5% CO2 in air. Efficiency of replating was used to measure degree of injury. Both halothane and isoflurane enhance the sensitivity of the RPAEC monolayers to injury by H2O2. The sensitizing effect of halothane was reversed by removing the anesthetic. Halothane and isoflurane thus enhance RPAEC sensitivity to injury by both H2O2 and PMA-stimulated neutrophils. In increasing RPAEC sensitivity to injury by oxygen metabolites, halothane and isoflurane may be inhibiting processes involved in intracellular antioxidant defenses.(ABSTRACT TRUNCATED AT 400 WORDS)

Barbiturate anesthetics inhibit thromboxane-, potassium-, but not angiotensin-induced pulmonary vasoconstriction.

Administration of the oxidant lipid peroxide tertiary butyl hydroperoxide (t-bu-OOH) in the isolated rabbit lung leads to acute pulmonary vasoconstriction, which is caused by the synthesis of thromboxane. The inhalational anesthetics, halothane, nitrous oxide, and cyclopropane, markedly enhance t-bu-OOH-induced pulmonary vasoconstriction and thromboxane production. The effects of the intravenous (iv) barbiturates thiopental and pentobarbital on t-bu-OOH-induced vasoconstriction were studied. Thiopental completely and pentobarbital partially blocked t-bu-OOH-induced vasoconstriction. Thiopental inhibited t-bu-OOH-induced synthesis of thromboxane and prostacyclin but pentobarbital did not. This inhibitory action of thiopental may be due to its antioxidant properties because similar inhibition has been observed of t-bu-OOH-induced thromboxane production with the antioxidants, vitamin E, or butylated hydroxylanisole. Thiopental and pentobarbital also inhibited the vasoconstriction induced by a thromboxane analog, epoxymethano prostaglandin H2 (U46619). Finally, both barbiturates partially inhibited the pulmonary vasoconstriction caused by potassium chloride, which requires calcium entry, but they did not inhibit the constriction caused by angiotensin II, which does not require calcium entry. These results suggest that pentobarbital and thiopental may block pulmonary vasoconstriction by inhibiting calcium entry.

Mepacrine attenuates pulmonary vasoreactivity in rabbits.

The organic peroxide tert-butyl hydroperoxide (t-bu-OOH) induces pulmonary vasoconstriction by stimulating production of thromboxane in the rabbit lung, possibly by activating phospholipase A2. t-bu-OOH-induced vasoconstriction and thromboxane production is augmented by inhalational anesthetic agents, perhaps due to an effect of anesthetic agents on membrane lipids. To further investigate the mechanism of thromboxane generation, we studied the influence of the phospholipase A2 inhibitor, mepacrine, in a dose known to inhibit the enzyme in other systems, on t-bu-OOH-induced pulmonary arterial vasoconstriction. We found that 10(-4) M mepacrine completely inhibited t-bu-OOH-induced vasoconstriction. We also found that mepacrine inhibited arachidonic acid-induced pulmonary vasoconstriction but did not inhibit thromboxane productions. We also investigated the effect of mepacrine on two other pulmonary vasoconstrictors, angiotensin II (ANG II) and KCl, which do not act through arachidonic acid metabolites in the rabbit lung. Mepacrine inhibited both ANG-II and KCl-induced vasoconstriction. The inhibition by mepacrine of pulmonary vasoconstriction is reversible if the drug is washed out of the lung. This effect of mepacrine cannot be explained by phospholipase inhibition alone and is consistent with prevention of smooth muscle contraction.

Inhalation anesthetics augment oxidant-induced pulmonary vasoconstriction: evidence for a membrane effect.

Inhalational anesthetics "fluidize" biologic membranes. Since arachidonate metabolism also occurs in cell membranes, anesthetic agents may modify arachidonic acid mediator production. The authors used the isolated perfused rabbit lung preparation to examine the effects of inhalational anesthetics on the production of arachidonate mediators. The oxidant tert-butyl-hydroperoxide (t-bu-OOH) is known to cause pulmonary vasoconstriction by causing increased production of thromboxane A2 (TxA2). The authors administered three anesthetics (halothane, cyclopropane, and nitrous oxide) of widely different potencies, at different dosages, each to three different groups of preparations and challenged the lungs at each anesthetic dose with t-bu-OOH. They found a dose-related augmentation of the pulmonary vasopressor response to t-bu-OOH. Preparations given t-bu-OOH alone showed no change in response over time. Lungs perfused with indomethacin (5 micrograms . ml-1 in Krebs-Henseleit buffer), ventilated with cyclopropane (2 MAC), and challenged with t-bu-OOH showed almost complete inhibition of the response to t-bu-OOH. Indomethacin at this concentration is a specific inhibitor of cyclooxygenase. The authors also have demonstrated significantly increased perfusate levels of thromboxane B2 (TxB2), the inactive metabolite of TxA2, after oxidant challenge during exposure to 2% halothane compared with TxB2 levels before halothane exposure. The authors believe that the augmented pressor response and mediator production occur because of increased substrate (arachidonic acid) availability induced by anesthetic agent.

Decreased sensitivity to metocurine in patients with upper motoneuron disease.

Responses to the nondepolarizing muscle relaxant, metocurine, were studied in eight hemiplegic and eight unmatched patients with normal motor strength during the general anesthetic given for various neurosurgical operations. Metocurine, 0.3 mg/kg, was administered intravenously, and indirectly evoked thumb twitch tensions were measured on both sides in the hemiplegic patients, and on one side in the normal patients. Arterial blood samples were obtained as twitch tension was recovering, and serum metocurine concentrations were determined using a specific radioimmunoassay. Percentage of paralysis was plotted as a function of log [metocurine] and the data were compared by analysis of covariance. For the normal motor strength patients, r = 0.84; for the unaffected arm of the hemiplegic patients, r = 0.69; and for the affected arm of the hemiplegic patients, r = 0.86, all significant at P less than 0.001. The mean plasma metocurine concentrations at 20, 25, 50, 75, and 80% paralysis were significantly different for all groups (P less than 0.001). The regression lines, in turn, did not overlap and were significantly different, each from the other (P less than 0.005). We were, however, unable to detect any significant deviation from parallelism among the three regression lines. We also measured time to 50% return of single twitch height for each data group as follows (mean +/- SEM: for NMS patients, 242 +/- 73 min; for the unaffected arm of hemiplegic patients, 116 +/- 60 min; and for the affected arm of hemiplegic patients, 59 +/- 36 min. By ANOVA and the Bonferroni test, each value was different from the other at P less than or equal to 0.01.(ABSTRACT TRUNCATED AT 250 WORDS)