Artificial surfactant in preterm rabbits with and without respiratory distress syndrome: difference of in vitro and in vivo activities

Sixty-eight rabbit fetuses of 27 days gestation were tracheotomised, artificially ventilated, and their lung mechanics studied in a body-enclosing plethysmograph. The animals were treated by tracheal instillation of natural surfactant concentrate or large unilamellar vesicles containing dipalmitoyl-phosphatidylcholine: egg phosphatidylglycerol, 91. Both preparations were highly surface active in terms of film adsorption and surface tension-lowering potential. Before treatment, the lung mechanics were analysed to indicate the presence of respiratory distress syndrome (RDS). Controls received 0.15 M saline. Differences were found between the in vitro and in vivo activities of both preparations in some animals. In 30 preterm animals with partial lung maturity and without respiratory distress syndrome, no significant effect could be achieved with either the natural or the artificial surfactant. In 38 animals with severe RDS, the tidal volume and compliance increased markedly within 15 min of substitution of both preparations. Compliance increased to 178% of the initial value in ventilated, control animals, to 391% in animals treated with natural, and to 344% in animals treated with artificial surfactant.Abbreviations RDS respiratory distress syndrome - DPPC dipalmitoylphophatidylcholine - TV tidal volume - C_rs compliance of the respiratory system     

Surfactant proteins and stable microbubbles in tracheal aspirates of infants with respiratory distress syndrome: relation to the degree of respiratory failure and response to exogenous surfactant

 Surfactant proteins (SP-A and SP-BC), albumin (ALB), and stable microbubble (SM) count were measured in tracheal aspirates from infants with respiratory distress syndrome (RDS) receiving single-dose Surfactant-TA (surfactant group, n = 32) or no surfactant (control group, n = 12), and those without RDS (non-RDS group, n = 8) to determine biochemical and biophysical status of surfactant in the course of RDS after surfactant replacement. Surfactant therapy resulted in immediate and sustained elevations of SP-BC/ALB and SM count with a rapid fall in ventilatory index to levels measured in the non-RDS group, whereas these indices improved slowly in the control group. The SP-A/ALB was initially low in both RDS groups and increased to levels measured in the non-RDS group by age 48 h. Multiple regression analysis showed that SP-BC/ALB, postnatal age, SM count, SM count/SP-A plus SP-BC, and surfactant therapy were independently associated with the severity of RDS as assessed by ventilatory index (r = 0.75, P < 0.0001; number of samples = 256). Infants with a relapse response to surfactant (n = 9) had levels of SP-A/ALB and SP-BC/ALB similar to those measured in the sustained group (n = 23), but had significantly lower SM count and SM count/SP-A plus SP-BC between 24 and 96 h of age. Conclusion Surfactant therapy normalizes the sur factant and respiratory status of infants with RDS. Surfactant dysfunction rather than depletion may explain the relapse response seen in some surfactant recipients. Received: 23 October 1995 / Accepted: 20 May 1996     

Bronchopulmonary dysplasia in infants with respiratory distress syndrome in a developing country: A prospective single centre-based study

The aim of this prospective study was to determine the incidence of bronchopulmonary dysplasia (BPD) in and the outcome of neonates ventilated for respiratory distress syndrome (RDS). The study was conducted in a developing country prior to the use of surfactant replacement therapy and the results are compared to published reports from the developed world. BPD was defined as oxygen dependency beyond day 28 of life. The incidence of BPD over a 9-month-period was 8.2% of all neonates requiring ventilation (n=169) and 41% (n=38) of neonates ventilated for RDS (n=92). Of those neonates who developed BPD, 26% were still being ventilated on day 28. Of the infants, 21 (55%) developed type 1 BPD and 17 (45%) type 2 BPD. There was no statistical difference in the severity of lung disease on any of the study days between type 1 and type 2 BPD although neonates with type 2 BPD required assisted ventilation and supplemental oxygen for a longer period: 30 versus 12 days and 95 versus 49 days, respectively. Of those neonates who developed BPD, 8 (21%) died prior to discharge from hospital and a further 5 infants (17%) died subsequent to discharge. Of the latter five, three died from treatable causes (gastroenteritisn=2, pneumonian=1). Of the 25 (83%) children seen at follow up, 68% were developing normally, 20% were classified as having suspect development and 12% had developed cerebral palsy at corrected postnatal ages of 12–24 months. None of the results differed significantly from those of neonates being ventilated in the developed world, except for the causes of post-discharge deaths.     

Resonance frequency in respiratory distress syndrome

AIM: To observe how the resonance frequency changes with the course of respiratory distress syndrome (RDS), by examining the effect of changing static compliance on the resonance frequency in premature infants. METHODS: In 12 ventilated premature infants with RDS (mean gestational age 26.6 weeks, mean birth weight 0.84 kg), resonance frequency and static compliance were determined serially using phase analysis and single breath mechanics technique respectively in the first seven days of life. RESULTS: The minimum number of measurements done in any one baby was three and maximum was five in this seven day study period. The first measurement in each baby was done within the first 72 hours of life. The increase in compliance in this period varied from 27% to 179%. The variation in the corresponding resonance frequency was within 2 Hz in eight babies and within 6 Hz in all recruited babies. CONCLUSIONS: The resonance frequency of the respiratory system in preterm infants with RDS remains remarkably constant in the early stages of the illness, despite relatively large changes in static compliance.     

Comparison between clinical and radiological classifications of infants with the respiratory distress syndrome (RDS)

Clinical and radiological classifications of the severity of the respiratory distress syndrome (RDS) were made in 55 infants. According to the clinical classification 17 infants belonged to the first class (mild RDS), 22 to the second (moderate RDS), and 16 to the third class (severe RDS). In the classification based on radiological findings the numbers of infants in classes 1, 2 and 3 were 18, 19 and 18 respectively. On the basis of both the clinical and radiological findings, 11 infants belonged to the mild RDS class, 11 to the moderate, and 12 to the severe RDS class. Thus, 34 infants had the same clinical and radiological classification. In 21 infants there were discrepancies between the clinical and the radiological classifications, but only one infant with the most severe radiological findings belonged to the mild RDS class and only one infant with mild radiological findings belonged to the worst RDS class.     

Precision of continuous neonatal ventilator respiratory mechanics is improved with selected optimal respiratory cycles

Given their high apparent variability, bedside continuous respiratory mechanics (RM) parameters [excepting tidal volume (V _T)] remain infrequently used for adjustment of neonatal ventilatory settings. RM parameters provided by ventilator (VRC) from ten recordings of newborns [10 min in synchronised intermittent mandatory ventilation and 10 min in assist/control (A/C)] were compared to those computed from visually selected assisted leak-free optimal respiratory cycles (SRC). Mean values, variability and ability to distinguish patients were compared between VRC and SRC. Dynamic resistances were more correlated (r ² = 0.95) than compliances (r ² = 0.42). V _Ts were correlated only in A/C (r ² = 0.78). C20/C was significantly higher in VRC (1.81 ± 0.67) than in SRC (1.23 ± 0.36) and frequently out of neonatal reference range. In A/C ventilation, V _T was higher in VRC (5.6 ± 1.8 ml/kg) than in SRC (4.8 ± 1.0 ml/kg) (p < 0.05). Displayed V _Ts do not reflect those found in optimal assisted breaths and therefore have incomplete value in assessing adequacy of ventilator settings. The variability of RM parameters provided by the ventilator is large, and coefficients of variation were significantly lower with optimal respiratory cycles (for resistance, compliance, V _T and C20/C; 27%, 26%, 18%, 24% in SRC and 36%, 35%, 40% and 33% in VRC). Selecting optimal cycles yields RM with two to three times higher discriminating power between patients. Conclusion: Current ventilator’s RM parameters have limited clinical use. Using optimal breaths to calculate RM parameters improves precision and discriminating power. For integration to ventilatory care, automation of this selection must be implemented first.     

Intraesophageal pressure monitoring and the severity of respiratory distress syndrome

The intraesophageal pressure difference (P _es) was measured in 46 preterm infants, 36 with respiratory distress (RDS). It was 12.7±2.9 mm Hg (mean±SEM) in severe RDS, 9.6±3.2 mm Hg in moderate RDS, and 5.2±1.6 mm Hg in infants with mild disease, measured at the age of 8 h. These intergroup differences were highly significant (P<0.001). During the first 96 h of life the P _es decreased to 6.1±1.1 mm Hg in severe RDS and to 4.8±1.2 mm Hg in moderate RDS. These changes were also highly significant (P<0.001) when compared with respective initial values. In infants with mild disease and in healthy preterm infants, the P _es did not alter significantly during the first 96 h of life.The correlations between gastric aspirate L/S ratio and the P _es at the age of 8 h (r=-0.65, P<0.001) and at the age of 24 h (r=-0.73, P<0.001) were highly significant.It is concluded that intraesophageal pressure monitoring is a valuable tool in the intensive care evaluation of the severity of RDS.Supported by grants from the Foundation for Paediatric Research and from the Medical Research Council, Academy of Finland     

Lung mechanics in premature infants with severe respiratory distress syndrome

Serial pulmonary function tests were performed in 13 preterm infants with severe RDS and 16 premature neonates with healthy lungs (8 intubated because of hypovention after birth, 8 were breathing spontaneously). Airflow was measured by a pneumotachograph, pressure changes were determined by airway pressure in ventilated infants or esophageal pressure in spontaneously breathing neonates. Pulmonary mechanics were calculated by a computerized system (PEDS/Medical Associated Services, Hatfield, Pennsylvania). In mechanically ventilated infants with RDS dynamic compliance of the respiratory system was low during the first 6 days. On Day 7-10 there were higher values for infants who could be extubated within the first 10 days compared to those, who had to be ventilated for more than 10 days. In the course of the disease, improvement in gas exchange preceded increase of compliance. Intraindividual comparisons in the acute and recovery phase of RDS and in infants with normal lungs showed higher values for compliance and lower values for airway pressure and resistance during spontaneous breaths. The differences between dynamic compliance of the respiratory system measured during mechanical ventilation, and dynamic lung compliance recorded during spontaneous breaths are due to influences of the respirator on the infant's lung.     

Thrombomodulin serum levels in ventilated preterm babies with respiratory distress syndrome

A soluble form of thrombomodulin (TM), an anticoagulant proteoglycan of the endothelial cell membrane, considered a marker of vascular endothelial damage, was measured in plasma of preterm infants with respiratory distress syndrome (RDS). In these patients, lung immaturity leads to endothelial leak of plasma proteins and to surfactant inhibition. In 18 babies with RDS, plasma TM concentration was significantly elevated compared with values of a matched group of babies without pulmonary disease (276.1 ng/ml vs 141.3 ng/ml) (P<0.05). Furthermore, TM levels of mechanical ventilated babies (IPPV) with severe RDS were higher than those of babies with moderate RDS and treated with nasal CPAP (340.9 ng/ml vs 174.2 ng/ml) (P<0.05). Conclusion These data show that TM can be used as marker of pulmonary endothelial damage in preterm babies treated with mechanical ventilation for RDS and suggest early intervention with exogenous surfactant to limit alveolar protein leakage and surfactant inactivation. Received: 20 February 1997 and in revised form: 7 July 1997 / Accepted: 8 July 1997     

Ventilator assessment of respiratory mechanics in paediatric intensive care

Many modern “paediatric” mechanical ventilators have in-built features for estimation of respiratory mechanics which could be useful in the management of ventilated infants and children. The aim of this study was to determine if such measurements were reproducible and accurate. Ventilator (Draeger Evita 4) displayed compliance (Cvent) and resistance (Rvent) values were assessed and compared to the results of respiratory system mechanics (respiratory system compliance (Crs) and resistance (Rrs)) measurements obtained using a single breath occlusion technique. Seventeen children (median age 5.1; range 0.3 to 16 yrs) were studied on 24 occasions. The mean coefficients of variations for the techniques were similar (Cvent 13%; Crs 11%; Rvent 16%; Rrs 14%). The mean (SD) Crs (22.8 (12.3) ml/cmH₂O) did not differ significantly from Cvent (22.1 (12.7) ml/cm H₂O) but the mean Rrs 21.0 (12.7) cmH₂O/l/s was significantly higher than the mean Rvent 32.0 (32.0) cmH₂O/l/s (p = 0.03). Bland and Altman analysis demonstrated a mean difference of −10.94 cmH₂O/l/s (SD 24.1) between Rrs and Rvent; the agreement between Rrs and Rvent decreased as Rrs increased (p = 0.008). Conclusions: Ventilator assessment of compliance, but not resistance, using the Evita 4 is reproducible and reliable. Funding note: Dr Harikumar was supported by the Wellcome Trust.     

Respiratory hospitalizations and respiratory syncytial virus prophylaxis in special populations

Palivizumab utilization, compliance, and outcomes were examined in infants with preexisting medical diseases within the Canadian Registry Database (CARESS) to aid in developing guidelines for potential “at-risk” infants in the future. Infants who received ≥1 dose of palivizumab during the 2006–2010 respiratory syncytial virus (RSV) seasons at 29 sites were recruited and utilization, compliance, and outcomes related to respiratory infection/illness (RI) events were collected monthly. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated for premature infants ≤35 completed weeks gestational age (GA) who met standard approval criteria (group 1) compared to those with medical disorders (group 2) using Cox proportional hazards regression models with adjustment for potential confounding factors. Of 7,339 registry infants, 4,880 were in group 1 and 952 in group 2, which included those with Down syndrome (20.3%), upper airway anomalies (18.7%), pulmonary diseases (13.3%), and cystic fibrosis (12.3%). Group 2 were older at enrolment (10.2 ± 9.2 vs. 3.5 ± 3.1 months, p < 0.0005), had higher GA (35.9 ± 6.0 vs. 31.0 ± 5.4 weeks, p < 0.0005), and were less compliant with treatment intervals (69.4% vs. 72.6%, p = 0.048). A greater proportion of group 2 infants were hospitalized for RI (9.0% vs. 4.2%, p < 0.0005) and RSV (2.4% vs. 1.3%, p = 0.003) (unadjusted). Being in group 2 was associated with an increased risk of RI (HR = 2.0, 95%CI 1.5–2.5, p < 0.0005), but not RSV hospitalization (HR = 1.6, 95%CI 0.9–2.8, p = 0.106). In infants receiving palivizumab, those with underlying medical disorders, though not currently approved for prophylaxis, are at higher risk for RI events compared with preterm infants. However, risk of RSV hospitalizations is similar.     

两种无创呼吸支持治疗极低出生体重儿呼吸窘迫综合征疗效分析

目的 探讨早期应用经鼻间歇正压通气（NIPPV）与加温湿化经鼻导管高流量通气（HHHFNC）治疗极低出生体重呼吸窘迫综合征（RDS）早产儿的临床疗效。方法 89例极低出生体重RDS早产儿依据初始呼吸支持治疗方式随机分为NIPPV组（46例）和HHHFNC组（43例）。分析两组患儿初始治疗失败率、肺表面活性物质（PS）使用率、呼吸支持治疗相关数据及各种并发症的发生率。结果 NIPPV组治疗72 h内气管插管率、PS使用率、有创辅助通气时间、无创辅助通气时间及总用氧时间、严重呼吸暂停发生率、肺炎发生率与HHHFNC组的差异无统计学意义（P > 0.05）;NIPPV组支气管肺发育不良、颅内出血、早产儿视网膜病、坏死性小肠结肠炎、动脉导管未闭、气漏的发生率与HHHFNC组的差异均无统计学意义（P > 0.05）;NIPPV组鼻损伤的发生率高于HHHFNC组（P < 0.05）。结论 对于极低出生体重RDS早产儿的初始呼吸支持治疗,HHHFNC与NIPPV疗效相当,是可供临床选择的另一种安全和有效的无创辅助通气治疗方式。     

Surfactant replacement therapy in neonatal respiratory distress syndrome

We conducted a prospective, randomized, controlled trial comparing the efficacy of two doses of a reconstituted bovine surfactant (Surfactant TA) in premature infants requiring mechanical ventilation shortly after birth for respiratory distress syndrome. Forty-six infants weighing 1000–1499 g were randomized into two groups: a low-dose group (23 infants given a single dose of 60 mg surfactant lipid/kg) and a high-dose group (23 infants given a single dose of 120 mg/kg). The mean (SD) age at which surfactant was given was 5.5 (±1.2) h in the low-dose group and 6.0 (±1.5) h in the high dose group. Both treatments improved oxygenation (increased arterial-alvcolar PO₂ ratio) with decreased mean airway pressure, the high-dose surfactant having a more beneficial effect in prolonging the response. Infants in the high-dose group had significantly less (P<0.05) incidence of both intraventricular haemorrhage and bronchopulmonary dysplasia. This prospective trial documents that a greater benefit can be obtained by increasing the dose of surfactant (120 mg/kg) beyond 60 mg/kg in the treatment of premature infants with severe respiratory distress syndrome (RDS).Abbreviations RDS respiratory distress syndrome - Surfactant TA Surfactant Tokyo-Akita - 5 kDa 5 kilodaltons - ELISA enzyme-linked immunosorbent assay - PDA patent ductus arteriosus - a/APO₂ arterial-alveolar oxygen tension ratio - MAP mean airway pressure - ^99mIc-DTPA 99m technetium diethylenetriamine pentacetate     

Respiratory outcome in preterm ventilated infants: importance of early respiratory system resistance

Since severity of acute respiratory distress syndrome (RDS) has been decreasing, the aetiology of long-term respiratory problems may have changed in preterm infants. We investigated whether early neonatal respiratory resistance (Rrs) and compliance (Crs) were important determinants of respiratory morbidity and lung function abnormalities during the 1st year of life in ventilated preterm infants. In 70 infants of less than 37 weeks gestation, mechanically ventilated within 24 h after birth, Rrs and Crs were assessed daily during the first 3 days of life and medians were calculated subsequently (Rrs_neo and Crs_neo). Rrs and Crs were reassessed 1 year later in 57 of 70 infants (Rrs_{1 year} and Crs_{1 year}). After correction for endotracheal tube size, increased Rrs_neo was significantly related to respiratory morbidity during the 1st year of life (OR 1.20, 95% CI 1.04 to 1.38; P =0.013), increased Rrs_{1 year} (multiplicative per kPa/l·s 1.017, 95% CI 1.000 to 1.034; P =0.045), and decreased Crs_{1 year} (multiplicative per kPa/l·s 0.986, 95% CI 0.974 to 0.998; P =0.023). Associations were not adversely affected by degree of prematurity, neonatal disease and treatment. Crs_neo did not show any significant associations. In the current surfactant era, increased Rrs_neo, and not Crs_neo was associated with poor respiratory outcome during the 1st year of life in preterm ventilated infants. Relationships were not adversely affected by measures reflecting degree of prematurity, neonatal disease, and ventilator conditions. Conclusion: our results indicate that inborn properties of the respiratory system have become increasingly important in relation to respiratory outcome instead of neonatal disease and treatment conditions.Abbreviations CLD chronic lung disease of the neonate - Crs respiratory system compliance - Crs _neo median respiratory system compliance of the first 3 days of life - Crs _{1 year} respiratory system compliance at 1 year of age - CV coefficient of variance - ERS European Respiratory Society - FiO₂ inspiratory oxygen fraction - PIP peak inspiratory pressure - RDS respiratory distress syndrome - Rrs respiratory system resistance - Rrs _neo median respiratory system resistance of the first 3 days of life - Rrs _neoC Rrs_neo corrected for endotracheal tube size - Rrs _{1 year} respiratory system resistance at 1 year of age - SBT single breath occlusion technique     

Cord blood Clara cell protein CC16 predicts the development of bronchopulmonary dysplasia

Clara cell protein (CC16) is an anti-inflammatory protein and a biomarker of pulmonary epithelial cells and alveolocapillary membrane injury in adults. We investigated whether low cord blood concentrations of CC16 are associated with the development of respiratory distress syndrome (RDS) and bronchopulmonary dysplasia (BPD) in preterm infants and the relationship between CC16 and its pro-inflammatory counterpart, the secretory phospholipase A₂ (sPLA₂) enzyme. CC16 concentration, sPLA₂ activity and IL-6 concentration were measured in cord blood plasma from 79 preterm infants (25 controls, 37 infants who developed RDS and 17 infants who developed BPD). After adjustment for gestational age and Apgar score at 5 min, the CC16 concentration was lower in BPD infants than in preterm controls (p<0.01). sPLA₂ activity was similar in all groups and the IL-6 concentrations were increased in both RDS and BPD infants (p<0.01 and p<0.05, respectively, vs. controls). We conclude that low cord blood CC16 concentrations in preterm infants independently predict the development of BPD. Low CC16 levels may reflect early lung injury, which contributes to the severity of RDS and progress towards BPD. Future studies are needed to assess whether the early administration of recombinant human CC16 in preterm infants with low cord blood CC16 prevents the development of BPD. This study was supported by grant 920-03-083 from the Netherlands Organisation for Scientific Research (NWO).     

Stable microbubble test on tracheal aspirate for the diagnosis of respiratory distress syndrome

BACKGROUND: Exogenous surfactant should be used as early as possible in the presence of respiratory distress syndrome (RDS), but diagnosis may only become clear late in the course of the disease. The stable microbubble test (SMT) in the tracheal aspirates could help in the decision to give early surfactant to preterm babies with respiratory distress. OBJECTIVES: The objective of this study was to evaluate the accuracy of the SMT on tracheal aspirate for the diagnosis of RDS in newborns requiring mechanical ventilation. METHODS: The test was performed on specimens obtained from 74 infants requiring mechanical ventilation, through routine suctioning. RESULTS: Patients with RDS and meconium aspiration syndrome (MAS) had a significantly lower stable microbubble count than non-RDS and non-MAS patients. Preterm infants without RDS had a significantly higher microbubble count than preterm babies with RDS and a similar count to that of term babies. Considering a cutoff point of 120 microbubbles/mm(2) for the diagnosis of RDS, the sensitivity of the microbubble test was 96.3% (95% CI: 79.1-99.8) and the specificity 97.6% (95% CI: 85.9-99.9). CONCLUSIONS: The SMT on tracheal aspirates is accurate for RDS diagnosis and may be useful to support the decision to give surfactant to newborns on mechanical ventilation.     

Effects of prolonged rupture of membranes on the development of respiratory distress syndrome in twin pregnancy

ABSTRACT. One hundred and ten twin infants between 26 to 36 weeks maturity admitted to a neonatal unit over a 3 1/2-year period were studied for the effects of prolonged rupture of membranes (PROM) on the development of respiratory distress syndrome (RDS). Fifty-eight infants were born following PROM, 24 developed RDS; in contrast 34 of the 52 infants who did not have PROM developed RDS. There was a marked difference of incidence of RDS between the presenting infants who had PROM and the non-presenting twins of the same pregnancy with intact gestational sacs (P<0.0005). Among the presenting infants, those with PROM also had significantly lower incidence of RDS compared with those without (P<0.005). This natural and paired controlled study has demonstrated a protective effect of PROM on the development of RDS.     

Supranormal Cardiac Output in the Dopamine- and Dobutamine-Dependent Preterm Infant

To evaluate the incidence of low cardiac output in preterm infants with respiratory distress syndrome (RDS), we measured cardiac output, stroke volume, heart rate, mean arterial blood pressure, and systemic vascular resistance at 8–48 hours of age in 30 preterm infants with RDS who were dependent on inotropic support. We then compared them to 23 normotensive preterm infants with RDS and 27 preterm infants without RDS. RDS infants had a higher cardiac output and lower systemic vascular resistance and blood pressure than infants without RDS. Infants treated with dopamine and dobutamine had a higher cardiac output and heart rate than infants on dopamine alone or the normotensive controls but a lower blood pressure and systemic vascular resistance than the normotensive controls. Supranormal cardiac output (>400 ml/min/kg) was detected in 57% of the infants in the dopamine + dobutamine subgroup (p= 0.009) versus 17% in the normotensive RDS subgroup and 12% in the dopamine subgroup. These data show that high cardiac output is relatively common in infants with RDS dependent on dopamine and dobutamine but is not reflected in the blood pressure.     

Antenatal ambroxol treatment does not prevent the respiratory distress syndrome in premature infants

The effectiveness of ambroxol in the prevention of neonatal respiratory distress syndrome and in reducing the need for intermittent mandatory ventilation and oxygen therapy was studied in 88 mothers whose infants was born between 24 and 34 weeks of gestation and who were randomized either for treatment with ambroxol (group A = 42) or served as control (group B = 46). There were no significant differences in the mean gestational age, birth weight or Apgar score between the two groups. We found no significant differences in occurrence of respiratory distress syndrome (55% vs 45%), in support by intermittent mandatory ventilation (71% vs 72%) or oxygen therapy (74% vs 75%) at 12 h of age between groups A and B. Conclusion This study does not suggest the efficacy of antenatal ambroxol treatment both for the prevention of neonatal respiratory distress syndrome and for the reduction of its severity. Received: 12 October 1995 and in revised form: 30 August 1996 / Accepted: 10 September 1996