Intensive Care Med (1999) 25: 76±80 Ó Springer-Verlag 1999
I. Schulze-Neick H. Werner D.J. Penny V. Alexi-Meskishvili P.E. Lange
Received: 26 March 1998 Final revision received: 9 June 1998 Accepted: 30 July 1998
)
I. Schulze-Neick ( ) × H. Werner × P. E. Lange Abteilung Angeborene Herzfehler, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, D-13 353 Berlin, Germany D.J. Penny Department of Pediatric Cardiology Royal Brompton & Harefield NHS Trust, Sydney Street, London SW3 6NP, UK V. Alexi-Meskishvili Abteilung Herz-Thorax-Gefäû-Chirurgie, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, D-13 353 Berlin, Germany Mailing address: Department of Pediatrics, Royal Brompton Hospital, Sydney Street, London SW3 6NP, UK email:
[email protected] Tel. + 44±171±351 8546 Fax + 44±171±351 8545
NE ON ATA L AN D P ED IATRI C IN TE NSI VE CA R E
Acute ventilatory restriction in children after weaning off inhaled nitric oxide: relation to rebound pulmonary hypertension
Abstract Objective: To assess the interaction between pulmonary hemodynamics and respiratory mechanics during acute pulmonary hypertension. Patients: Ventilated and paralysed children treated with inhaled nitric oxide because of post-operative pulmonary hypertension. Interventions: Weaning of inhaled nitric oxide. Measurements: Air flow and airway pressure, calculation of dynamic respiratory system compliance and respiratory system resistance for each breath by multiple linear regression. Results: In four patients, increases in pulmonary arterial pressure from 26.1 to 56.7 mmg (p < 0.001) during weaning off nitric oxide were associated with decreases in tidal volume (from 9.7 ® 8.2 ml/kg, p < 0.01) and reductions in dynamic respiratory system compliance (from 0.52 ® 0.34 cmH20/ml/kg,
Introduction After surgery for congenital heart defects associated with left-to-right shunting, children are at risk of pulmonary arterial hypertension and of profound, episodic increases in pulmonary vascular resistance (pulmonary hypertensive crises) in the early post-operative period [1]. Although, in most patients, increases in pulmonary vascular resistance are effectively treated with inhaled nitric oxide [2], in some, rebound pulmonary hypertension and pulmonary hypertensive crises are observed when inhaled nitric oxide is withdrawn [3].
p < 0.001), while respiratory system resistance was unchanged. Conclusions: Impaired ventilation during acute pulmonary hypertension is predominantly related to a reduction in respiratory system compliance. Key words Nitric oxide × Rebound × Pulmonary hypertension × Dynamic respiratory system compliance ´ Interaction ´ Congenital heart disease
These rebound pulmonary hypertensive crises appear to show a clinically similar picture to acute pulmonary hypertensive crises in the untreated patient. Characteristically, they are associated with increases in pulmonary arterial pressure, reductions in systemic arterial pressure and signs of impaired cardiac output. Furthermore, as in the untreated patient, they are associated with an inability to maintain alveolar ventilation and the clinical impression of ªstiffº lungs [4], suggesting altered respiratory mechanics. In order to obtain further insights into the relationships between pulmonary hypertensive crises and respi-
77
Table 1 Cardiovascular and ventilatory parameters with and without inhaled nitric oxide therapy (PAP mean pulmonary artery pressure, MAP mean arterial pressure, O2Sat haemoglobin oxygen saturation (%), TV tidal volume (ml/kg), PIP peak inspiratory pressure (cmH2O), Cdyn dynamic respiratory system compliance
(ml/cmH2O/kg), Rrs respiratory system resistance (cmH2O/ml/s); **p < 0.01 and ***p < 0.001 during inhaled nitric oxide (NO) therapy versus during rebound pulmonary hypertension (Re-PHT))
PAP mm Hg
MAP mm Hg
O2Sat %
TV (ml/kg)
PIP (cmH2O)
Cdyn (ml/cmH2O/kg)
Rrs (cmH2O/ml/s)
During NO Mean Range
26.3 (22±29)
59.3 (57±65)
95.3 (92±98)
9.7 (7.5±11.2)
23.5 (21±25)
0.52 (0.36±0.62)
0.0992 (0.084±0.117)
During Re-PHT Mean Range p -value
56.7 (52±60) ***
56.5 (54±61) NS
88.5 (82±94) **
8.2 (5.7±10.0) **
23.8 (21±25) NS
0.34 (0.26±0.45) ***
0.1008 (0.086±0.117) NS
ratory mechanics, we studied lung function during weaning off inhaled nitric oxide in 11 patients. We report data from four in whom the discontinuation of nitric oxide therapy was associated with significant rebound pulmonary hypertension.
Patients and methods All four patients were young infants (0.41±0.83 years; 4.1±6.1 kg) who underwent surgery for congenital heart defects associated with ventricular left-to-right shunting. In the early post-operative period, all remained nasally intubated, with cuffed endotracheal tubes (internal diameter 3.0±3.5 mm), underwent pressure-limited ventilation (rate 22±28/min, I:E-ratio = 1:2, PEEP 2±3 mmg) and had indwelling lines to monitor systemic arterial, central venous and pulmonary arterial pressures. All demonstrated evidence of increased pulmonary vascular resistance in the early post-operative period, with pulmonary arterial pressures exceeding 50 % of systemic arterial pressures, and signs of reduced cardiac output. All were treated with sedation, hyperventilation and paralysis. Inhaled nitric oxide was commenced at concentrations measured as 10±20 parts per million (ppm) (chemiluminescence, ECO-Physics, Druenten, Switzerland). Inhaled nitric oxide therapy was conducted according to a protocol approved by the ethical commission of our institution, and all parents had given informed consent. Once haemodynamic stability was achieved and therapeutic as well as ventilatory variables were gradually and reliably reduced, the patient was weaned off nitric oxide according to clinical routine. This was performed by stepwise reduction of the nitric oxide concentration to 5 and 3 ppm, which was then discontinued. However, as discontinuation of nitric oxide resulted in rebound pulmonary hypertension in these patients, the treatment was reinstituted. All were successfully weaned during subsequent attempts using lower discontinuation doses of nitric oxide. During the weaning period, pulmonary and systemic arterial pressures were recorded at 1 min intervals. Lung function was monitored continuously, and neither ventilatory (avoiding rate dependent compliance changes) nor other treatment variables were changed. Handling of the patient was avoided, while the only variable modified in the measurement setup was the nitric oxide concentration. Air flow was continuously measured at the opening of the endotracheal tube, using a calibrated double hot wire anemometer (NVM1 monitor, Bear Medical Systems, Riverside, Calif., USA). Airway pressure was measured by a manometer (Isotech,
Healthdyne Cardiovascular Inc. Philadelphia, Pa., USA) connected via a T-piece to the endotracheal tube connector. Flow and pressure signals were digitized at 100 hz (DT2801-A, Data Translation, Marlboro, Mass., USA), stored on computer and processed offline, using the ANADAT software (Jason Bates, Meakins Christie Laboratory, Montreal, Canada), with flow signals being integrated to derive volume. The equation of motion for a single-compartment lung model was solved using multiple linear regression [5], and total respiratory system (rs) resistance (R) and dynamic compliance (Cdyn) were then computed using the equation Ç /s + EEP Paw = Ers*V + Rrs*V where Paw is airway pressure, Ers is elastance ( = 1/compliance), V Ç /s is flow, and EEP is end-expiratory pressure reflectis volume, V ing auto-PEEP or inadvertant PEEP [6, 7]. Single breaths were discarded when the coefficient of determination was less than 0.95. We analyzed the whole respiratory cycle, since this resulted in the lowest standard deviations of measured variables [7]. Standard statistical tests (ANOVA) were used to compare the indices of lung function before rebound, at peak pulmonary arterial pressure during rebound and again after recovery.
Results The standard deviation for all measured variables was less than 3 % during inhaled nitric oxide ventilation. However, during rebound pulmonary hypertension, tidal volume, compliance and resistance changed with each breath, while airway pressures remained unchanged. Rebound pulmonary hypertension after the discontinuation of nitric oxide therapy resulted in an increase in pulmonary arterial pressure (p < 0.001) and a decrease in arterial oxygen saturation (p < 0.01) (Table 1). These were associated with a progressive decrease in tidal volume (p < 0.01) without a rise in EEP, while airway pressures and flow-volume relationships of the breaths remained unchanged (pressure limited ventilation) (Fig. 1). As a result, there was marked reduction in respiratory system compliance(p < 0.001), while respiratory system resistance was unchanged (Figs. 2, 3), thus showing the picture of acute ventilatory
78
Fig. 1 Acute reduction in tidal volume during rebound pulmonary hypertension. Pressure-volume loops (ml of tidal volume versus airway pressure in mm H2O) in a single patient during rebound pulmonary hypertension demonstrating the gradual decrease in tidal volume in the presence of pressure-limited ventilation
restriction. Furthermore, there was a close statistical relationship between instantaneous increases in pulmonary arterial pressure and reductions in respiratory system compliance (r2 = 0.81). All hemodynamic and ventilatory parameters returned to baseline within minutes of the reinstitution of nitric oxide therapy (Fig. 4).
Discussion This is the first study to describe the breath-by-breath changes in respiratory mechanics during acute pulmonary hypertension in children. It demonstrates that an acute pulmonary hypertensive crisis after weaning off inhaled nitric oxide is associated with marked reductions in dynamic respiratory system compliance, which is promptly reversed when the nitric oxide treatment is reinstituted. Although obtained in a small patient number, the findings were highly significant and highlight both the close inter-relationship of pulmonary haemodynamics and lung function as well as the syndrome of ªstiff lungsº associated with pulmonary hypertensive crisis in children after cardiac surgery. We discontinued nitric oxide treatment at concentrations of 3±5 ppm, as was commonplace during the early use of nitric oxide therapy [8]. However, with further experience in recent years, lower concentrations before discontinuation are now used.
Fig. 2 Acute change of compliance during rebound pulmonary hypertension. Values for dynamic respiratory system compliance (Cdyn, ml/cmH2O/kg) and respiratory system resistance (Rrs, cmH2O/ml/s) are displayed for each breath together with the corresponding pulmonary arterial pressure (PAP, mmHg) during inhaled nitric oxide (NO) therapy and rebound pulmonary hypertension after attempted weaning off NO
The close interaction of pulmonary haemodynamics and lung function is well recognised in experimental studies. Increases in left atrial pressure as a result of the transient obstruction of pulmonary venous return in dogs was well correlated with reductions in pulmonary compliance [9, 10]. In contrast, varying pulmonary arterial flow did not cause significant changes in respiratory mechanics [11]. Pulmonary hypertensive crises are associated with increases in pulmonary arterial resistance, possibly increases in pulmonary venous resistance, as well as reductions in pulmonary blood flow and pulmonary capillary blood volume. The identification of the primary haemodynamic precipitant for the changes in respiratory mechanics which we observed will require further study. The predominant effect of the pulmonary hypertensive crisis on respiratory mechanics we observed was a reduction in dynamic respiratory system compliance, causing acute ventilatory restriction. Since we observed neither increasing end-expiratory pressures during these periods nor release of trapped air as an increase in tidal volume after these episodes, our findings cannot be explained by dynamic hyperinflation and a shift upward to a less compliant portion of the pressure-volume curve [12]. Our observations differ from those of Schindler et al. [13], where a marked increase in respiratory system re-
79
Fig. 3 Change of pulmonary mechanics during rebound PHT. Group data for the relationship between proportional changes in mean pulmonary artery pressure and respiratory mechanics. Each point represents the average of a 30 s recording of lung function and the corresponding values of mean pulmonary artery pressure
sistance with a minor change in static compliance was found to be associated with spontaneous pulmonary hypertensive episodes in children. The apparent lack of agreement with our data may be explained by a number of differences in study design. Firstly, we continuously observed respiratory mechanics throughout the development of rebound pulmonary hypertension and terminated this process with the reinstitution of nitric oxide therapy at an early stage when systemic hypotension was about to develop. In contrast, measurements in Schindlers study were performed ªas soon as possibleº after the onset of pulmonary hypertensive crisis after suctioning of the endotracheal tube. Both the time delay caused by this and the suctioning itself may have caused increased airway resistance by allowing the accumulation of interstitial edema caused by prolonged severe pulmonary vascular congestion [9]. Secondly, there are also differences in measurement technique. While we measured dynamic respiratory system compliance, Schindler et al. measured static changes. Differences in static and dynamic compliance measurement may reflect the visco-elastic properties of the lung tissue [7]. These visco-elastic properties of the lungs are high in infants and are attributed to increased tissue density in these patients, and will be further increased in conditions involving increased lung fluid. Finally, our finding of unchanged respiratory system resistance may point to the fact that we observed a different pathophysiology of pulmonary hypertension (rebound versus spontaneous), which was linked to the early development of acute ventilatory restriction [14] rather than the later
Fig. 4 Recovery of acutely changed respiratory mechanics with NO. Group data for the values for mean pulmonary artery pressure (PAP, mmg), dynamic respiratory system compliance (Cdyn, ml/cmH2O/kg) and respiratory system resistance (Rrs, cmH2O/ml/ s) in all patients during nitric oxide (NO) therapy, during NO weaning followed by rebound pulmonary hypertension, and after the reinstitution of NO therapy
development of ventilatory obstruction. However, it is clear that, irrespective of the precise changes in respiratory mechanics associated with these episodes, they may have important implications for the propagation of these crises with abnormalities in respiratory mechanics leading to progressive hypoxemia and hypercapnia, which will feed the pulmonary hypertension in an iterative fashion. In summary, our findings demonstrate the dramatic effects of rebound pulmonary hypertension on dynamic respiratory mechanics. Our data suggest that reductions in respiratory system compliance resulting in acute ventilatory restriction explain the finding of ªstiff lungsº associated with these episodes.
80
References 1. Hopkins RA, Bull C, Haworth SG, De Leval MR, Stark J (1991) Pulmonary hypertensive crisis following surgery for congenital heart defects in young children. Eur J Cardiothorac Surg 5(12):628±634 2. Haydar A, Malhere T, Mauriat P, Journois D, Pouard P, Denis N, Lefebvre D, Safran D, Vouhe P (1992) Inhaled nicric oxide for postoperative pulmonary hypertension in patients with congenital heart defects. Lancet 340(Dec 19/ 26):1545 3. Atz AM, Adatia I, Wessel DL (1996) Rebound pulmonary hypertension after inhalation of nitric oxide. Ann Thorac Surg 62(6):1759±1764 4. von Basch S (1887) Über eine Funktion des Capillardruckes in den Lungenalveolen. Wiener Med Blätter 15: 465 5. Seear M, Wensley D, Werner H (1991) Comparison of three methods for measuring respiratory mechanics in ventilated children. Pediatr Pulmonol 10: 291±295
6. Bates J, Shardonofsky F, Stewart D (1989) The low-frequency dependance of respiratory system resistance and elastance in normal dogs. Respir Physiol 78: 369±382 7. Lanteri CJ, Kano S, Nicolai T, Sly PD (1995) Measurement of dynamic respiratory mechanics in neonatal and pediatric intensive care: the multiple linear regression technique. Pediatr Pulmonol 19: 29±45 8. Rossaint R, Falke KJ, Lopez F, Slama K, Pison U, Zapol WM (1993) Inhaled nitric oxide for the adult respiratory distress syndrome. N Engl J Med 238: 399±405 9. Borst HG, Berlund E, Whittenberger JL, Mead J, McGregor M, Collier C (1957) The effect of pulmonary vascular pressures on the mechanical properties of the lungs of anesthetized dogs. J Clin Invest 36: 1708±1714 10. Gray BA, McCaffree DR, Sivak ED, McCurdy HT (1978) Effect of pulmonary vascular engorgement on respiratory mechanics in the dog. J Appl Physiol Respir Environ Exercise Physiol 45(1):119±127
11. Uhlig T, Wildhaber JH, Eber E, Sly PD (1997) Vagal reflex is not responsible for changes in airway and lung tissue mechanics due to vascular engorgement in young piglets. Pediatr Res 42: 533±537 12. Volgyesi G, Webster P, Tremblay L, Slutsky A (1998) Apparent decrease in respiratory system compliance during methacholine challenge is caused by gas trapping. Am J Respir Crit Care Med 157:A106 13. Schindler MB, Bohn BJ, Bryan AC, Cutz E, Rabinovitch M (1995) Increased respiratory system resistance and bronchial smooth muscle hypertrophy in children with acute postoperative pulmonary hypertension. Am J Respir Crit Care Med 152: 1347±1352 14. Cotes JE (1993) Lung function in disease. Blackwell, London Edinburgh Boston Melbourne Paris Berlin, pp 517±519