339
Journal of Neonatal-Perinatal Medicine 6 (2013) 339–344 DOI 10.3233/NPM-1372713 IOS Press
Original Research
Volume targeted ventilation and arterial carbon dioxide in extremely preterm infants S. Shah∗ and A. Kaul Aditya Birla Memorial Hospital, Chinchwad, Pune, India
Received 01 March 2013 Revised 22 September 2013 Accepted 27 September 2013
Abstract. AIM: The aim of our study was to quantify arterial carbon dioxide levels (PaCO2 ) achieved by ventilating extremely preterm neonates in volume guarantee mode targeting tidal volumes of approximately 4 ml/kg. METHODS: We performed a prospective trial on preterm infants with gestational age ≤28 weeks, birth-weight ≤1000 grams, postnatal age 6 episodes requiring stimulation in 6 hours or >1 episode requiring positive pressure ventilation, an arterial Ph 60 mmHg, or treatment with oxygen more than 50%. Infants intubated in the delivery room were ventilated using T piece resuscitator during transfer to NICU and then switched over to volume guarantee plus synchronized intermittent positive pressure ventilation (SIPPV + VG mode). All infants were ventilated using SIPPV + VG mode with Babylog 8000 plus ventilator (Draeger Inc, Germany). Initial ventilator settings were: set tidal volume at 4 ml/kg of body weight, positive end-expiratory pressure at 5 cms H2 O, inspiratory time 0.35 seconds and backup ventilator rate of 40/minute. The PIP limit (Pmax) was set at 25 cm H2 O during initiation of ventilation. If the set tidal volume was not delivered then the Pmax was increased by 5 cm·H2 O to a maximum of 35 cm·H2 O. Measured PIP, delivered tidal volume, measured ventilation rate, measured minute ventilation were recorded every hourly and corresponding to arterial blood gases. When the leak around endotracheal tube was more than 30%, then one size bigger endotracheal tube was inserted. Endotracheal tube was cut at a distance of 4 cms from the angle of mouth. Standard neonatal ventilatory tubing sets supplied with the ventilator were used and no additional gadgets
S. Shah and A. Kaul / Volume ventilation and CO2
such as capnograph were used in the ventilator circuit. All blood gases were taken from indwelling arterial lines. Blood gas analysis was done within 1 hour of commencement of ventilation and subsequently at the clinician’s discretion. If the infant had hypocarbia, then the set tidal volume was decreased by 1 ml/kg. If the infants had hypercarbia, then the set tidal volume was increased by 1 ml/kg. Subsequently such infants were removed from the study and analysis was done only for pCO2 levels corresponding to a set tidal volume of 4 ml/kg. Data were collected on infants for the first 48 hours during the time they were ventilated on volume guarantee mode. Surfactant treatment, extubation and reintubation criteria were not mandated. Morphine was used on a PRN basis as a sedation policy. Clinical care was offered as per the unit protocol. Primary outcome was the average first PaCO2 in neonates, ventilated in volume guarantee mode from admission to the neonatal intensive care and number of these infants with an unacceptable PaCO2 on the first blood gas. Pre-specified secondary outcomes were the average PaCO2 in infants ventilated in the volume guarantee mode during the first 48 hrs of ventilation and number of blood gases with an unacceptable PaCO2 in the first 48 hours. Hypocarbia was defined as PaCO2 60 [7]. Neonates with hypocarbia or hypercarbia were classified as having unacceptable PaCO2 . 2.2. Statistical analysis Statistical analysis was performed using SPSS 10.0 software. For all the continuous variables which were normally distributed, mean and standard deviation were calculated. Linear regression analysis was done
341
Table 1 General characteristics of the study population Characteristics
Mean ± SD
Range
Gestational age (weeks) Birth weight (grams) Ventilation (hours)
27.0 ± 1.6 877 ± 130 25 ± 16
24–28 weeks 450–1000 grams 8–48 hours
to look for relationship between targeted tidal volume, measured minute ventilation and PaCO2 levels.
3. Results From Jan 2010 to Dec 2012, a total of 92 infants were eligible for enrolment. Sixty two infants were enrolled while thirty were excluded from the study (refusal of consent-26, inability to ventilate using SIPPV + VG mode – 2, inability to secure to an indwelling arterial line-2). Reasons for ventilation were: 80% (50/62) respiratory distress syndrome (all received at least one dose of surfactant), 15% (9/62) prematurity without lung disease, 5% (3/62) apnea of prematurity. A total of 62 infants were enrolled and 218 blood gases recorded up to 48 hours of age. The mean gestational age and standard deviation was 27.08 ± 1.6 weeks (range 24–28 weeks). The mean birth weight and standard deviation was 877.28 ± 130 grams (range 450–1000 grams). Thirty two infants were intubated and ventilated in the delivery room while 30 infants were ventilated after a failed trial of CPAP. The analysis of first arterial blood gas and corresponding ventilator settings for all ventilated infants are shown in Table 2. The analysis of all blood gases and corresponding ventilatory settings for the first 48 hours are also shown in Table 2. The number of PaCO2 values at various cut-off levels of hypocarbia
Table 2 Summary of ventilation parameters and PaCO2 levels for all infants during the first arterial blood gas measurement and during the first 48 hours of ventilation Ventilation and blood gas parameters
Mean ± SD (Range) during the first Mean ± SD (Range) during the first arterial blood gas measurement 48 hours of ventilation
Delivered tidal volume (ml/kg) Measured minute ventilation (ml/kg) MAP (cms H2 O) mPIP (cms H2 O) Measured ventilatory rate PaCO2 values pH
4.30 ± 0.60 (3–5.1) 254 ± 48 (111–400) 9.4 ± 0.7 (6–14) 16.4 ± 4.1 (6.2–25) 55 ± 7 (36–62) 39.5 ± 6.3 (25–55) 7.33 ± 0.08 (7.22–7.58)
4.17 ± 0.48 (3–5.1) 250.13 ± 42.5 (111–400) 7.6 ± 0.5 (6–15) 11.4 ± 4.1 (6.1–25) 53 ± 6 (35–62) 39.37 ± 6.9 (25–55) 7.31 ± 0.07 (7.20–7.58)
342
S. Shah and A. Kaul / Volume ventilation and CO2
Fig. 1. Number of infants or blood gas values corresponding to various PaCO2 cut off levels First blood gas (n = 62). All blood gases (n = 218).
and hypercarbia are shown in Fig. 1 for both analysis groups.
4. Discussion In our study, we show that if extremely preterm infants are ventilated using Drager Babylog 8000 + ventilator in VG mode at a tidal volume setting of 4 ml/kg, the mean PaCO2 in the first 48 hours was 39.37 ± 6.99 mmHg. The mean PaCO2 on the first arterial blood gas for infants ventilated in VG mode from admission to the nursery was 39.5 ± 6.3. PaCO2 levels were in the acceptable range in 93% of infants at the time of first ABG measurement. PaCO2 levels were in the acceptable range in 91% of blood gases during the first 48 hours of ventilation. We have not made any attempt to compare VG mode with any another mode of ventilation. We just wanted to show the PaCO2 values achieved with VG at a set tidal volume of 4 ml/kg. We chose a fairly homogenous group of extremely low birth weight infants since this group is very different from larger preterm infants. All infants were ventilated using SIPPV + VG mode thus supporting all spontaneous breaths and partially eliminating the effect of spontaneous breathing rate on CO2 elimination. Cheema et al. [8] first showed the feasibility of volume guarantee ventilation in a four hour crossover trial involving 40 infants. The VG group was able to achieve equivalent ventilation using lower peak pressures. Abubakar and Keszler [9] compared effect the effect of SIPPV + VG versus SIMV + VG in a short cross over trial. Infants in the SIPPV + VG group had less variability in tidal volume and decreased work of
breathing as compared to SIMV + VG group. Another crossover trial by Scopesi and colleagues [10] suggested that in weaning phase, volume guarantee in modes that support every breath may be more effective than a combination of SIMV plus volume guarantee. Herrera et al. [11] showed a reduction in ventilatory support using SIMV + VG compared to SIMV alone in a short crossover trial. Peak pressures were further reduced and spontaneous breaths increased (suggestive of increased work of breathing) when the target tidal volume was reduced from 4.5 ml/kg to 3 ml/kg. Lista et al. [12] measured inflammatory mediators in tracheal aspirates of preterm infants with RDS ventilated using VG mode using 2 different target tidal volumes (3 ml/kg and 5 ml/kg). They showed that when target volume of 3 ml/kg is used inflammatory cytokines are increased. This could be due to atelectrauma resulting from inadequate delivery of tidal volume. In a randomized controlled trial Cheema and colleagues [13] evaluated initial arterial blood gases in infants with RDS shortly after they were placed on SIPPV mode versus SIPPV + VG mode (VG 4 ml/kg). Analysis revealed that infants >25 weeks gestation had significantly less hypocarbia when volume guarantee was used (27% vs 61%, p < 0.05). However, all 4 infants between 23–25 weeks had out of range CO2 (target range 37.5–52.5 mmHG) using VG mode of ventilation. The numbers are small but further study is needed in this group of micropremies. Dawson and Davies [4] performed a retrospective review of 50 newborns who were ventilated using synchronized intermittent mandatory ventilation in volume guarantee mode with a targeted tidal volume of 4 ml/kg and concluded that severe hypo or hypercarbia could be avoided over 90% of the time. In another retrospective review Nassabeh-Montazami et al. [14] studied the effect of instrumental dead-space on volume guarantee ventilation and obtained normative data for initial tidal volume associated with normocapnia in ELBW infants. The tidal volume/kg needed for normocapnia was inversely related to weight. Mean tidal volume/kg of infants 700 g (P < 0.001). Clinical studies have so far used a target tidal volume of 3–6 ml/kg and majority have studied larger preterm infants. Differences between the delivered and set tidal volume are important as the tidal volume influences blood gas exchange. Although the set tidal volume was 4 ml/kg, the delivered tidal volume varied between
S. Shah and A. Kaul / Volume ventilation and CO2
Fig. 2. Scatter plots with linear regression equations for PaCO2 versus measured tidal volume and PaCO2 versus measured minute ventilation are shown in Figs. 1 and 2. Both the scatter plots show a decreasing trend of PaCO2 values with increasing tidal volumes and minute ventilation. PCO2 —Linear (PCO2 ).
3–5.1 ml/kg. This is the due to changes in the respiratory mechanics (compliance and resistance) and the ventilator trying to target the set tidal volume as per the inbuilt algorithm. Clinicians cannot assume that a preset tidal volume will give adequate gas exchange and hence it is important to monitor CO2 levels. The average measured VT was above set VT and PIP was low (Table 2). Most babies were breathing above set VT as seen also in Fig. 2. The mean pH shows mild metabolic acidosis. One could argue that some of the infants may have been receiving insufficient support leading to increased work of breathing. However, the measured ventilator rate was within the normal range (35–62) during the duration of ventilation. Hence it seems unlikely that the infants were receiving insufficient support. We also recorded the measured minute ventilation (MMV) and correlated it with PaCO2 values in the first 48 hours. The measured minute ventilation gives a better reflection of the ventilatory input than the measured tidal volume. The mean measured minute ventilation was 250.13 ml/kg and standard deviation 42.5 ml/kg. This is comparable to values reported in the literature [15]. Both the scatter plots show a decreasing trend of PaCO2 values with increasing tidal volumes and minute ventilation. MMV, which is what the ventilator calculates is an indirect measure of the efficiency of CO2 removal. Alveolar Minute ventilation (AMV) is what really matters and that does not always have a linear relationship
Fig. 3.
343
PCO2 —Linear (PCO2 ).
with MV. Rapid shallow breathing has a high dead space to VT ratio and results in lower AMV. This could be the reason for the wide scatter in our plots. However, determining AMV is not feasible in clinical practice and MMV serves as a surrogate marker for alveolar ventilation [16]. A drawback of our study is that we did not perform continuous monitoring of measured minute ventilation and measured tidal volume, but recorded it intermittently. However, this reflects usual clinical practice in most of the units. Another obvious limitation is lack of control group. Many studies have investigated the link between arterial carbon dioxide levels and neurodevelopmental and respiratory outcomes of premature infants. Hypocarbia, particularly PaCO2
Journal of Neonatal-Perinatal Medicine 6 (2013) 339–344 DOI 10.3233/NPM-1372713 IOS Press
Original Research
Volume targeted ventilation and arterial carbon dioxide in extremely preterm infants S. Shah∗ and A. Kaul Aditya Birla Memorial Hospital, Chinchwad, Pune, India
Received 01 March 2013 Revised 22 September 2013 Accepted 27 September 2013
Abstract. AIM: The aim of our study was to quantify arterial carbon dioxide levels (PaCO2 ) achieved by ventilating extremely preterm neonates in volume guarantee mode targeting tidal volumes of approximately 4 ml/kg. METHODS: We performed a prospective trial on preterm infants with gestational age ≤28 weeks, birth-weight ≤1000 grams, postnatal age 6 episodes requiring stimulation in 6 hours or >1 episode requiring positive pressure ventilation, an arterial Ph 60 mmHg, or treatment with oxygen more than 50%. Infants intubated in the delivery room were ventilated using T piece resuscitator during transfer to NICU and then switched over to volume guarantee plus synchronized intermittent positive pressure ventilation (SIPPV + VG mode). All infants were ventilated using SIPPV + VG mode with Babylog 8000 plus ventilator (Draeger Inc, Germany). Initial ventilator settings were: set tidal volume at 4 ml/kg of body weight, positive end-expiratory pressure at 5 cms H2 O, inspiratory time 0.35 seconds and backup ventilator rate of 40/minute. The PIP limit (Pmax) was set at 25 cm H2 O during initiation of ventilation. If the set tidal volume was not delivered then the Pmax was increased by 5 cm·H2 O to a maximum of 35 cm·H2 O. Measured PIP, delivered tidal volume, measured ventilation rate, measured minute ventilation were recorded every hourly and corresponding to arterial blood gases. When the leak around endotracheal tube was more than 30%, then one size bigger endotracheal tube was inserted. Endotracheal tube was cut at a distance of 4 cms from the angle of mouth. Standard neonatal ventilatory tubing sets supplied with the ventilator were used and no additional gadgets
S. Shah and A. Kaul / Volume ventilation and CO2
such as capnograph were used in the ventilator circuit. All blood gases were taken from indwelling arterial lines. Blood gas analysis was done within 1 hour of commencement of ventilation and subsequently at the clinician’s discretion. If the infant had hypocarbia, then the set tidal volume was decreased by 1 ml/kg. If the infants had hypercarbia, then the set tidal volume was increased by 1 ml/kg. Subsequently such infants were removed from the study and analysis was done only for pCO2 levels corresponding to a set tidal volume of 4 ml/kg. Data were collected on infants for the first 48 hours during the time they were ventilated on volume guarantee mode. Surfactant treatment, extubation and reintubation criteria were not mandated. Morphine was used on a PRN basis as a sedation policy. Clinical care was offered as per the unit protocol. Primary outcome was the average first PaCO2 in neonates, ventilated in volume guarantee mode from admission to the neonatal intensive care and number of these infants with an unacceptable PaCO2 on the first blood gas. Pre-specified secondary outcomes were the average PaCO2 in infants ventilated in the volume guarantee mode during the first 48 hrs of ventilation and number of blood gases with an unacceptable PaCO2 in the first 48 hours. Hypocarbia was defined as PaCO2 60 [7]. Neonates with hypocarbia or hypercarbia were classified as having unacceptable PaCO2 . 2.2. Statistical analysis Statistical analysis was performed using SPSS 10.0 software. For all the continuous variables which were normally distributed, mean and standard deviation were calculated. Linear regression analysis was done
341
Table 1 General characteristics of the study population Characteristics
Mean ± SD
Range
Gestational age (weeks) Birth weight (grams) Ventilation (hours)
27.0 ± 1.6 877 ± 130 25 ± 16
24–28 weeks 450–1000 grams 8–48 hours
to look for relationship between targeted tidal volume, measured minute ventilation and PaCO2 levels.
3. Results From Jan 2010 to Dec 2012, a total of 92 infants were eligible for enrolment. Sixty two infants were enrolled while thirty were excluded from the study (refusal of consent-26, inability to ventilate using SIPPV + VG mode – 2, inability to secure to an indwelling arterial line-2). Reasons for ventilation were: 80% (50/62) respiratory distress syndrome (all received at least one dose of surfactant), 15% (9/62) prematurity without lung disease, 5% (3/62) apnea of prematurity. A total of 62 infants were enrolled and 218 blood gases recorded up to 48 hours of age. The mean gestational age and standard deviation was 27.08 ± 1.6 weeks (range 24–28 weeks). The mean birth weight and standard deviation was 877.28 ± 130 grams (range 450–1000 grams). Thirty two infants were intubated and ventilated in the delivery room while 30 infants were ventilated after a failed trial of CPAP. The analysis of first arterial blood gas and corresponding ventilator settings for all ventilated infants are shown in Table 2. The analysis of all blood gases and corresponding ventilatory settings for the first 48 hours are also shown in Table 2. The number of PaCO2 values at various cut-off levels of hypocarbia
Table 2 Summary of ventilation parameters and PaCO2 levels for all infants during the first arterial blood gas measurement and during the first 48 hours of ventilation Ventilation and blood gas parameters
Mean ± SD (Range) during the first Mean ± SD (Range) during the first arterial blood gas measurement 48 hours of ventilation
Delivered tidal volume (ml/kg) Measured minute ventilation (ml/kg) MAP (cms H2 O) mPIP (cms H2 O) Measured ventilatory rate PaCO2 values pH
4.30 ± 0.60 (3–5.1) 254 ± 48 (111–400) 9.4 ± 0.7 (6–14) 16.4 ± 4.1 (6.2–25) 55 ± 7 (36–62) 39.5 ± 6.3 (25–55) 7.33 ± 0.08 (7.22–7.58)
4.17 ± 0.48 (3–5.1) 250.13 ± 42.5 (111–400) 7.6 ± 0.5 (6–15) 11.4 ± 4.1 (6.1–25) 53 ± 6 (35–62) 39.37 ± 6.9 (25–55) 7.31 ± 0.07 (7.20–7.58)
342
S. Shah and A. Kaul / Volume ventilation and CO2
Fig. 1. Number of infants or blood gas values corresponding to various PaCO2 cut off levels First blood gas (n = 62). All blood gases (n = 218).
and hypercarbia are shown in Fig. 1 for both analysis groups.
4. Discussion In our study, we show that if extremely preterm infants are ventilated using Drager Babylog 8000 + ventilator in VG mode at a tidal volume setting of 4 ml/kg, the mean PaCO2 in the first 48 hours was 39.37 ± 6.99 mmHg. The mean PaCO2 on the first arterial blood gas for infants ventilated in VG mode from admission to the nursery was 39.5 ± 6.3. PaCO2 levels were in the acceptable range in 93% of infants at the time of first ABG measurement. PaCO2 levels were in the acceptable range in 91% of blood gases during the first 48 hours of ventilation. We have not made any attempt to compare VG mode with any another mode of ventilation. We just wanted to show the PaCO2 values achieved with VG at a set tidal volume of 4 ml/kg. We chose a fairly homogenous group of extremely low birth weight infants since this group is very different from larger preterm infants. All infants were ventilated using SIPPV + VG mode thus supporting all spontaneous breaths and partially eliminating the effect of spontaneous breathing rate on CO2 elimination. Cheema et al. [8] first showed the feasibility of volume guarantee ventilation in a four hour crossover trial involving 40 infants. The VG group was able to achieve equivalent ventilation using lower peak pressures. Abubakar and Keszler [9] compared effect the effect of SIPPV + VG versus SIMV + VG in a short cross over trial. Infants in the SIPPV + VG group had less variability in tidal volume and decreased work of
breathing as compared to SIMV + VG group. Another crossover trial by Scopesi and colleagues [10] suggested that in weaning phase, volume guarantee in modes that support every breath may be more effective than a combination of SIMV plus volume guarantee. Herrera et al. [11] showed a reduction in ventilatory support using SIMV + VG compared to SIMV alone in a short crossover trial. Peak pressures were further reduced and spontaneous breaths increased (suggestive of increased work of breathing) when the target tidal volume was reduced from 4.5 ml/kg to 3 ml/kg. Lista et al. [12] measured inflammatory mediators in tracheal aspirates of preterm infants with RDS ventilated using VG mode using 2 different target tidal volumes (3 ml/kg and 5 ml/kg). They showed that when target volume of 3 ml/kg is used inflammatory cytokines are increased. This could be due to atelectrauma resulting from inadequate delivery of tidal volume. In a randomized controlled trial Cheema and colleagues [13] evaluated initial arterial blood gases in infants with RDS shortly after they were placed on SIPPV mode versus SIPPV + VG mode (VG 4 ml/kg). Analysis revealed that infants >25 weeks gestation had significantly less hypocarbia when volume guarantee was used (27% vs 61%, p < 0.05). However, all 4 infants between 23–25 weeks had out of range CO2 (target range 37.5–52.5 mmHG) using VG mode of ventilation. The numbers are small but further study is needed in this group of micropremies. Dawson and Davies [4] performed a retrospective review of 50 newborns who were ventilated using synchronized intermittent mandatory ventilation in volume guarantee mode with a targeted tidal volume of 4 ml/kg and concluded that severe hypo or hypercarbia could be avoided over 90% of the time. In another retrospective review Nassabeh-Montazami et al. [14] studied the effect of instrumental dead-space on volume guarantee ventilation and obtained normative data for initial tidal volume associated with normocapnia in ELBW infants. The tidal volume/kg needed for normocapnia was inversely related to weight. Mean tidal volume/kg of infants 700 g (P < 0.001). Clinical studies have so far used a target tidal volume of 3–6 ml/kg and majority have studied larger preterm infants. Differences between the delivered and set tidal volume are important as the tidal volume influences blood gas exchange. Although the set tidal volume was 4 ml/kg, the delivered tidal volume varied between
S. Shah and A. Kaul / Volume ventilation and CO2
Fig. 2. Scatter plots with linear regression equations for PaCO2 versus measured tidal volume and PaCO2 versus measured minute ventilation are shown in Figs. 1 and 2. Both the scatter plots show a decreasing trend of PaCO2 values with increasing tidal volumes and minute ventilation. PCO2 —Linear (PCO2 ).
3–5.1 ml/kg. This is the due to changes in the respiratory mechanics (compliance and resistance) and the ventilator trying to target the set tidal volume as per the inbuilt algorithm. Clinicians cannot assume that a preset tidal volume will give adequate gas exchange and hence it is important to monitor CO2 levels. The average measured VT was above set VT and PIP was low (Table 2). Most babies were breathing above set VT as seen also in Fig. 2. The mean pH shows mild metabolic acidosis. One could argue that some of the infants may have been receiving insufficient support leading to increased work of breathing. However, the measured ventilator rate was within the normal range (35–62) during the duration of ventilation. Hence it seems unlikely that the infants were receiving insufficient support. We also recorded the measured minute ventilation (MMV) and correlated it with PaCO2 values in the first 48 hours. The measured minute ventilation gives a better reflection of the ventilatory input than the measured tidal volume. The mean measured minute ventilation was 250.13 ml/kg and standard deviation 42.5 ml/kg. This is comparable to values reported in the literature [15]. Both the scatter plots show a decreasing trend of PaCO2 values with increasing tidal volumes and minute ventilation. MMV, which is what the ventilator calculates is an indirect measure of the efficiency of CO2 removal. Alveolar Minute ventilation (AMV) is what really matters and that does not always have a linear relationship
Fig. 3.
343
PCO2 —Linear (PCO2 ).
with MV. Rapid shallow breathing has a high dead space to VT ratio and results in lower AMV. This could be the reason for the wide scatter in our plots. However, determining AMV is not feasible in clinical practice and MMV serves as a surrogate marker for alveolar ventilation [16]. A drawback of our study is that we did not perform continuous monitoring of measured minute ventilation and measured tidal volume, but recorded it intermittently. However, this reflects usual clinical practice in most of the units. Another obvious limitation is lack of control group. Many studies have investigated the link between arterial carbon dioxide levels and neurodevelopmental and respiratory outcomes of premature infants. Hypocarbia, particularly PaCO2
