Over the past decades, cardiac surgery for children with congenital heart disease has improved. Still these patients have high postoperative mortality, long stays in the intensive care unit, and prolonged ventilation periods.Reference Boneva, Botto, Moore, Yang, Correa and Erickson1 The aim of the present study is to improve knowledge about duration of postoperative ventilation; this is not only of interest to parents but is also important for surgical planning, taking into account the available intensive care beds. And it might also be helpful in the planning of intensive care beds and ventilation capacities. Risk predictive scoring systems are valid tools for benchmarking to compare the outcome of children in different hospitals or different populations. Furthermore, they help to detect problems and may lead to improvement.
The Risk Adjustment in Congenital Heart Surgery-1 (RACHS-1) score is a consensus-based score of six risk categories according to the complexity of the operation with rising postoperative mortality.Reference Jenkins, Gauvreau, Newburger, Spray, Moller and Iezzoni2 It has been confirmed by other studies.Reference Boethig, Jenkins, Hecker, Thies and Breymann3–Reference Simsic, Cuadrado, Kirshbom and Kanter5 As the length of postoperative ventilation is also related to the complexity of the operation, this study analyses whether the RACHS-1 score may also be a predictor for the length of postoperative ventilation in children after cardiac surgery. Several studies showed a relation between young age at time of operation and higher rates of postoperative complications and a higher mortality.Reference Polito, Patorno and Costello6–Reference Szekely, Sapi, Kiraly, Szatmari and Dinya8 One of those postoperative complications is prolonged postoperative ventilation. Padly et al.Reference Padley, Cole and Pye7 described 2011 that newborns <30 days had twice as long a stay in the intensive care unit, significant more complications and longer postoperative ventilation.
Secondary issues of the present study are the effects of patient’s age on the length of postoperative ventilation, changes of postoperative ventilation during the study time, and a comparison of the mortality in our study with that reported in the literature.
Methods
This retrospective study was performed in a paediatric cardiac intensive care unit in a tertiary university hospital after having gained permission from the local ethical committee (ekmed@uk-koeln.de, 02 August, 2017).
Over a period of 6 years (October, 2007–October, 2013), the data of 598 children with congenital heart disease who underwent cardiac surgery were identified.
Thirty-nine patients have been excluded because of sepsis or severe infection (n = 4), bronchopulmonary disease (n = 15), resuscitation and extracorporal membrane oxygenation (n = 9), prematurity (n = 2), emergency surgery (n = 6), and other severe disease that affected mortality or ventilation (n = 3). Five hundred and fifty-nine patients could be analysed for mortality. To analyse the postoperative ventilation, all deceased patients (n = 39) were excluded, 520 cases remained.
The collected demographic data included age, weight, gender, diagnosis, type of cardiac surgery, and length and mode of postoperative ventilation.
All patients were categorised by the RACHS-1 score. Mortality was defined as death in hospital or within 30 days after surgery. Five age groups were built based on the recommendations of the “International Conference of Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human use”(ICH) with a further division within the group of infants: newborns 1–27 days, infants 28 days – 6 months, infants >6 months – 2 years, children >2 years – 11 years, and adolescent >11 years – 18 years.
Ventilation modes
Ventilation modes for invasive ventilation synchronised intermittent mandatory ventilation (SIMV), biphasic positive airway pressure (BIPAP), airway pressure release ventilation (APRV), and high-frequency oscillation ventilation (HFOV) were used; some patients received inhaled nitric oxide (iNO) in addition. For non-invasive ventilation, nasal continuous positive airway pressure (nCPAP) and mask-CPAP with flow-triggered mode of pressure support ventilation (PSV) were used.
Extubation
Due to the great distance between operating theatre and intensive care unit, all patients were transferred with the tracheal tube in place. Weaning was then started as soon as cardiovascular function was stable. Extubation was performed after adequate ventilation with continuous positive airway pressure (CPAP) or moderate PSV for at least 1 hour according to the decision of the attending physician. The extubation failure was defined by reintubation within 12 hours, and for the analysis of the ventilation time, these short extubation times were counted as ventilation time.
Statistical analysis
All data were stored in an Excel (Microsoft®, Redmond, Washington, United States of America) spreadsheet and statistically analysed with SPSS 24 (SPSS®, IBM, Armonk, New York, United States of America). Postoperative duration of ventilation was described by survival curves. Differences between two survival curves were analysed with a Wilcoxon–Gehan test. Differences during the study time were analysed with a Mann–Whitney test and cross tables with Spearman’s Rho. Cox regression was used for a multifactorial analysis of age and RACHS-1 Score. p-values <0.05 were considered as statistically significant.
RACHS-1 group 5 contained only two patients; therefore, this group was excluded from analysis of ventilation times.
Results
Demographic data
After exclusion of patients with additional comorbidities (see Methods), 559 patients with cardiac surgery at the University Hospital Cologne could be analysed and classified with the RACHS-1 Score. Detailed data (age distribution, mortality) are shown in Table 1.
Table 1. Demographic data and mortality
At the time of surgery, about half of the patients were younger than 6 months, less than 25% were older than 2 years. Forty-four percent were female and 56% were male. Overall mortality was 7% (n = 39). Extubation failure was 7.1%.
Most operations fell within RACHS-1 group 3 (n = 218) and group 2 (n = 201).
RACHS-1 group 1 included older children (53.1% 2–11 years), RACHS-1 group 2 more infants (46% 1–5 month), RACHS-1 group 4 newborns (47.5%) and infants (30.5%), and RACHS-1 groups 5 and 6 mainly newborns (100%, respectively, 85.1%). In RACHS-1 group 3, all ages were almost equally represented.
Mortality
Overall mortality was 7%. In RACHS-1 group 1, mortality was 0%, in RACHS-1 group 2, 3.5% (n = 7), in RACHS-1 group 3, 6.4% (n = 14), RACHS-1 group 4, 10.2% (n = 6), in RACHS-1 group 5, 100% (n = 2), and in RACHS-1 group 6, 21.3% (n = 10).
RACHS-1 and ventilation
The median length of postoperative ventilation rose significantly with a higher RACHS-1 score (Table 2). RACHS-1 group 1 had a median duration of ventilation of 9 hours (IQR 6.8–13.0), RACHS-1 group 2 of 30 hours (IQR 11.5–85.4), RACHS-1 group 3 of 57 hours (IQR 12.9–134.6), RACHS-1 group 4 of 70 hours (IQR 29.4–164.5), and RACHS-1 group 6 of 188 hours (IQR 126.7–276.5). In a pairwise comparison of the survival curves (Fig. 1), all groups differ significantly except for groups 4 and 3 (p < 0.01).
Table 2. Ventilation times
* Total ventilation time = invasive + non-invasive ventilation times
Table 3. Cox regression
RACHS-1 group 6 and age 0–27 days are the reference groups for other RACHS-1 and age groups as they have the longest ventilation times
There was a high variance in the RACHS-1 groups 2 and 3; therefore, the most common operations in these RACHS-1 groups were analysed separately. Common operations were repair of ventricular septal defects (VSD) (n = 50), total cavo-pulmonary connections (TCPC) (n = 32), partial cavo-pulmonary connections (PCPC) (n = 40), and repair of tetralogy of Fallot (TOF) (n = 31). Patients undergoing TOF repair had a significantly longer duration of postoperative ventilation (median 97 hours, IQR 33.7–168.9) compared to the other patients in the same RACHS-1 group 2 (26.7 hours, IQR 11.0–68.1), while patients undergoing TCPC, PCPC, or VSD repair needed a similar or slightly shorter postoperative ventilation time than their RACHS-1 group.
Age and ventilation
The median length of postoperative ventilation also rose significantly with younger age. Newborns had a median ventilation time of 138 hours (IQR 72–226), babies from 1 to 5 months 56 hours (IQR 23–144), and infants from 6 to 23 months 31 hours (IQR 11–81). Older children (2–11 years) and adolescents (11–18 years) did not differ in the median ventilation time; they both needed a median ventilation of 11 hours (IQR 8–27).
Figure 1. Ventilation times in RACHS-1 groups.
Figure 2. Ventilation time in age groups.
Likewise, the survival curves of postoperative ventilation (Fig. 2) showed significant difference between the age groups (Wilcoxon–Gehan test, p < 0.001).
RACHS-1 group 3, which had the most balanced age distribution, showed significant difference between the age groups (p < 0.001) with median ventilation times for newborns of 121 hours (IQR 74–210), babies 94 hours (IQR 34–172), toddlers 43 hours (IQR 20–105), older children 14 hours (IQR 9–65), and adolescents 12 hours (IQR 9–27). A pairwise analysis between the groups revealed no significance between the two youngest and the two oldest age groups, but for all others.
Multifactorial analyses
The Cox regression (Table 3) showed a clear significance (p < 0.001) for the influence of RACHS-1 score and age on postoperative ventilation time. The hazard ratio for shorter ventilation times increased with a lower RACHS-1 group or an older age.
Changes during the study period
For detection of changes during the study period, the first 3 years had been compared to the second 3 years of the study period. Overall mortality in the second part of the study period was 5.8% compared to 8.3% in the first part of the study. This effect was present in all RACHS-1 groups. The most notable reduction in mortality between the two study periods was noticed in RACHS-1 group 6 with a mortality of 37.5% in the first study period and 12.9% in the second study period.
The median invasive ventilation time was significantly shorter in the second study period (48 hours, IQR 11–122, in the first period versus 16 hours, IQR 10–73, in the second period, p < 0.001). The use and the duration of non-invasive ventilation increased in the second study period. In the first study period, 19.8% received non-invasive ventilation and 59.6% in the second study period. Therefore, the difference in the total ventilation time did not change significantly (median 54 hours, IQR 12–128, in the first study period and 39 hours, IQR 12–119, in the second study period). Detailed data are listed in Table 2.
Discussion
This study confirms the RACHS-1 scoreReference Jenkins, Gauvreau, Newburger, Spray, Moller and Iezzoni2 in the sense that mortality increases with a higher RACHS-1 score. Comparing the mortality data of the Paediatric Cardiac Care Consortium (PCCC) and the hospital discharge data sets of the original RACHS-1 scoreReference Jenkins, Gauvreau, Newburger, Spray, Moller and Iezzoni2 with our results demonstrates similar mortality rates in RACHS-1 groups 1–3, but lower mortality in RACHS-1 group 4 with 10.2% compared to 19.2–19.6% and RACHS-1 group 6 with 21.3% compared to 41.5–47.0%. Other studies from DenmarkReference Larsen, Pedersen, Jacobsen, Johnsen, Hansen and Hjortdal4 and GermanyReference Boethig, Jenkins, Hecker, Thies and Breymann3 also reported higher mortality rates of 57 and 40.1% in RACHS-1 group 6. During the study periods, the mortality declined even more. In the second half of the study (2011–2013), the mortality in RACHS-1 group 6 was only 12.9%. Similar data are only reported in a study from JapanReference Nakayama, Shibasaki, Shime, Nakajima, Mizobe and Sawa9 analysing a period from 2005 to 2009 with a mortality rate of 11% in RACHS-1 group 6 and in the data of the Society of Thoracic Surgeons Congenital Heart Surgery Database with 16.0% in bigger centersReference Jacobs, Mayer and Mavroudis10; in the latter, they analysed the STAT-score, which is similar to the RACHS-1 score in the highest group (STAT group 5 is comparable to RACHS-1 group 6).
One reason for the low mortality in our study in the higher RACHS-1 groups could be the high number of performed operations in a university hospital and hence the expertise of the whole team including surgeons, anaesthesiologists, cardiologists, and intensive care team. Another reason is the general improvement of therapy in paediatric cardiac surgery and perioperative medicine.
RACHS-1 score may be used as a predictor for postoperative ventilation. Only few studies have analysed this subject before. Szekely et al.Reference Szekely, Sapi, Kiraly, Szatmari and Dinya8 compared a lower RACHS-1 score (groups 1–3) with a higher RACHS-1 score (groups 4–6) and found a correlation between higher RACHS-1 score and a medium (>61 hours) and long (>7 days) ventilation time. Likewise, Odek et al.Reference Odek, Kendirli and Ucar11 found a correlation between low RACHS-1 score and early extubation, though they included only RACHS-1 groups 1–3. The already mentioned Japanese studyReference Nakayama, Shibasaki, Shime, Nakajima, Mizobe and Sawa9 is the only other study that describes the postoperative ventilation time in relation to the RACHS-1 score. They showed Kaplan–Mayer analyses for postoperative mechanical ventilation, catecholamine levels, and intensive care unit (ICU) stays, which differed significantly (p < 0.05) in RACHS-1 groups except RACHS groups 4 and 6. Compared to our study, where we found significant differences between groups 4 and 6, they had fewer patients in these groups (57 versus 90 in our study, which corresponds to 9.8 versus 17.3%) that might explain the difference. Unfortunately, they did not provide figures on mean ventilation times for comparison.
The relationship between RACHS-1 score and postoperative ventilation time can be explained by more complex operations with an increased severity of trauma, longer duration of operation and cardiopulmonary bypass, higher rate of cases with complete circulatory arrest, more capillary leak, and overall higher risk for complications.
One exception in regard to postoperative ventilation was the repair of TOF. Those patients needed significant longer postoperative ventilation time with a huge variation (median 79 hours, IQR 34–169) than the rest of RACHS-1 group 2. Different studies have tried to find risk factors and explanations for long postoperative ventilation after repair of Fallot. Young age, long cardiopulmonal bypass time, genetic syndromes, and previous operations seemed to be some of the reasons.Reference Jochman, Atkinson, Quinonez and Brown12–Reference van Dongen, Glansdorp and Mildner14 The huge variance in ventilation duration might also be explained by the large variability in the anatomy of TOF, ranging from less severe forms of (sub)pulmonary obstruction to nearly pulmonary atresia. As the grade of pulmonary obstruction has significant implications on the methods of surgical repair and on residual lesions, particularly pulmonary valve regurgitation, ventilation times may also be influenced.
The STAT-ScoreReference Jacobs, O’Brien and Jacobs15,Reference O’Brien, Clarke and Jacobs16 is a newer mortality score that is not solely related to expert opinion like the RACHS-1 score but which is based on the database of the Society of Thoracic Surgeons (STS) and the European Association for Cardiothoracic Surgery (EACTS) congenital heart surgery database. The STAT score consists of five groups of different operations. An analysis of our data with that score showed an overall significant association between STAT score and ventilation times (p < 0.001, Wilcoxon–Gehan test), but looking into detail group 2 showed a longer median ventilation time than group 3 (no significant pairwise difference between these two groups, p = 0.84). This result is in line with the prolonged postoperative length of stay (PLOS) in the analysis by the STS. Median ventilation time in STAT group 1 was 16 hours (IQR 9–58), 51 hours (IQR 13–125) in STAT group 2, 33 hours (IQR 12–125) in STAT group 3, 103 hours (IQR 32–215) in STAT group 4, and median 169 hours (IQR 127–277) in STAT group 5. Although the STAT score is based on a better and objective database, it provides no advantage compared to the RACHS-1 score concerning the length of postoperative ventilation.
The length of postoperative ventilation or the rate of extubation failure also rises with a younger age which is in line with many other studies. Young children with congenital heart disease seem to have a different physiology, often they have low birth weight and recurrent infections; hence, they have a lower cardiorespiratory reserve and are predisposed for long postoperative ventilation.Reference Tabib, Abrishami, Mahdavi, Mortezaeian and Totonchi17,Reference Traiber, Piva and Fritsher18 Other studies found differences between premature babies (<37 weeks gestation and <30 days),Reference Rooney, Donohue and Bush19 newborns (<30 days),Reference Polito, Patorno and Costello6,Reference Padley, Cole and Pye7 or infants (<1 year)Reference Szekely, Sapi, Kiraly, Szatmari and Dinya8 to older children. This is confirmed by the present study that shows the effect of age on the postoperative ventilation in five age groups. It is striking that there was no significant difference between the group 2–11 years and the adolescent 12–17 years; hence, it can be concluded that age is a significant factor in postoperative ventilation times, especially for children under 2 years of age. There are more young patients in the higher RACHS-1 groups; nevertheless, the Cox regression shows that RACHS-1 is an independent factor for the duration of postoperative ventilation time. Likewise, the analysis of RACHS-1 group 3 with an almost equal distribution of age shows that younger age is a predictor for prolonged postoperative ventilation time. On the other hand, some congenital heart diseases depend on an operation at a certain age, so some operations and RACHS-1 groups are not independent of age.
Invasive postoperative ventilation was significantly shorter in the second study period, but the proportion of non-invasive ventilated patients was higher and the duration longer; therefore, the total ventilation time remained the same. There are many reasons in favour of non-invasive ventilation (spontaneous breathing reduces the intrathoracic pressure compared to mechanical breathing and thus the right ventricular afterload, which has a positive effect on blood flow, reduced need for analgetics and sedatives). These benefits are confirmed by other studies that show reduced work of breathing, better oxygenation, and low reintubation rates through non-invasive ventilation after cardiac surgery,Reference Zhang, Tan and Shi20,Reference Silva, Andrade, Maux, Bezerra and Duarte21 but our data show that this change in treatment did not necessarily shorten overall ventilation time.
Limitations
The major limitation of the present study is the retrospective study design. Due to the lack of corresponding electronically stored data after 2014, it was not possible to evaluate the data of the last 4 years.
Postoperative ventilation time is influenced by many factors, RACHS-1 score and age are just two of the most important ones, and the influence of other factors (e.g. length of cardiopulmonary bypass, length of cardiac arrest fluid management) cannot be excluded.
Although the RACHS groups and age are independent factors, both are linked in the form that certain operations are linked to a certain age.
The paediatric cardiac intensive care unit was established in 2007; hence, the first study period 2007–2010 describes the initial period of this ICU at Cologne University Hospital. Comparison between the two study periods does exclusively demonstrate not only the progress in paediatric cardiac intensive care medicine but also the improvements in a newly constructed intensive care unit.
Conclusion
The mortality rate in this study is comparable to data from other studies in RACHS-1 groups 1–3, but significantly lower in RACHS-1 groups 4 and 6. For the duration of postoperative ventilation time, the RACHS-1 score is also a strong predictor. Ventilation time increases with higher RACHS-1 groups, but there are exceptions in specific operations, especially in the repair of TOF, which shows a longer ventilation time with a large variance compared to the other operations within RACHS-1 group 2. Age is another independent risk factor for a longer postoperative ventilation time. Up to an age of 2 years, younger age is associated with longer postoperative ventilation times, while an age of 2–18 years shows no significant difference in the duration of postoperative ventilation time. The trend towards shorter invasive ventilation times and increased use of non-invasive ventilation does not shorten overall ventilation time.
These data provide better insight into ventilation times and allow better planning of operations in terms of available intensive care beds.
Acknowledgements
The authors thank Prof Brockmeier and Prof Emmel for their contribution to the planning of the study. The authors thank S.v.P. for improving the English writing and approving the final version.
Financial Support
This research did not receive any financial support.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guidelines (Professional Code of Conduct for the physicians of North Rhine-Westphalian/Germany) and with the Helsinki Declaration of 1975, as revised in 2008, and has been approved by the institutional committees (Ethical committee of the University of Cologne).
