Introduction

Diaphragmatic paralysis (DP) due to phrenic nerve injury after congenital cardiac surgery is an important respiratory complication resulting with respiratory insufficiency, lung infections, prolonged hospital stay time and even death [1–9]. It is accepted that procedures in congenital cardiac surgery are the major causes of DP with an incidence ranging from 0.3 to 12.8% [2, 10–12].

It should be considered as a severe and life-threatening respiratory complication especially in younger age. Diaphragm paralysis causes respiratory distress because of paradoxical motion of the affected diaphragm and contralateral shift of the mediastinum [3]. The common feature of DP after cardiac surgery in children is inability in weaning from mechanical ventilation or requirement of reintubation. This situation is the major cause of lung infections and related complications such as atelectasis and recurrent pneumonia [3–5]. Great majority of older children have the advantage of overcoming DP as they use their intercostal muscles for respiration when compared with infants [6–9].

Suspicion is the first step in diagnosis of DP. Among infants in whom there is no cardiac cause for failure to wean from mechanical ventilation, DP should be suspected. Elevated hemidiaphragm on chest X-rays of patients will lead the clinician to confirm DP with other diagnostic tools such as ultrasound (US), fluoroscopy and electromyography (EMG) [3, 13].

In children and especially in infants, early diaphragmatic plication provides treatment for DP with good results. In our study, we tried to determine the incidence, pathological features, clinical relevance and treatment of DP after pediatric cardiac surgery.

Patients and methods

Between August 1996 and January 2005, 3,071 patients underwent cardiac surgery for congenital heart disease. Postoperative DP was diagnosed in 152 (4.9%) patients. Diaphragmatic plication was performed in 42/152 patients (27.6%). Before the operation, the family of the patients were informed about the details and risks of the operation and written informed consents were obtained from the parents of all patients. After DP was diagnosed, this complication was discussed with the family in detail (the risks, treatment options). This study was approved by the Internal Review Board of our Institute.

Operative and postoperative clinical course data were collected from both computer-based databank system, surgery notes and patient files. All patients were followed at Baskent University Hospital at least once in 6 months for a detailed physical examination combined with echocardiographic assessment. Follow-up data were obtained from pediatric cardiology and cardiovascular surgery outpatient follow-up notes. Chest X-rays were reviewed at 1st, 6th and 12th month after discharge.

Diaphragmatic paralysis was suspected in children who failed to wean from mechanical ventilation or in those with persistent respiratory distress when there is no cardiac cause. Decreased respiratory sounds in auscultation, paradoxical breathing during spontaneous ventilation and elevated hemidiaphragm on chest X-rays led us to use fluoroscopy, US and/or EMG. There was no sign of DP in preoperative chest X-rays of any patient. When chest X-rays did not have a diagnostic value in patients with persistant respiratory distress; bilateral DP was suspected and immediate fluoroscopy or EMG was performed for diagnosis.

All the patients were ventilated with a ServoVentilator 300 (Siemens, Stockholm, Sweden) after the cardiac surgery. The initial mode of ventilation was pressure-regulated volume control. Once the patient was breathing spontaneously and ready for weaning, the ventilator mode was switched to either pressure or volume-controlled, synchronized, intermittent, mandatory ventilation. Usual criteria for extubation were adequate oxygenation on FiO2 of ≤0.4, maintenance of a pH >7.35, a \( P_{{{\text{CO}}_{2} }} \) of <45 mmHg on continuous positive airway pressure of ≤6 cm H2O with pressure support of ≤8 cm H2O, level of consciousness consistent with adequate airway protective reflexes, and acceptable hemodynamics. The attending pediatric intensivists (both a pediatric cardiologist and a cardiac surgeon) approved all planned extubations before proceeding. Corticosteroids were routinely administered before the extubation.

Indications for diaphragmatic plication were recurrent reintubations (n=22), failure to wean from ventilator (n=12), recurrent lung infections (n=5) and persistent respiratory distress (n=3). Respiratory distress was defined as: tachypnea, oxygen dependency or CO2 retention. The decision of respiratory distress was supported with arteriel blood gas analysis (Tachypnea, oxygen dependency, oxygen saturation <90%, pH <7.35, a \( P_{{{\text{CO}}_{2} }} \) of >45 mmHg was defined as CO2 retention. The patients with paralysed diaphragms but without any clinical symptoms or with normal arteriel blood gas analysis did not underwent plication.

Surgical technique

In all patients, diaphragmatic plications were performed using a thoracic approach, through the sixth or seventh intercostal space with lateral thoracotomy. Fixation of the diaphragm was performed by a series of non-absorbable longitudinal U sutures placed in a posterior to anterior direction and always reinforced at both ends over a Teflon buttress, as described by Schwartz and Filler [2]. Suturing process was performed by placing at least four loops of sutures.

Statistics

Statistical data was evaluated with SPSS (Statistical Package for the Social Sciences for windows, version 10.0, Chicago, IL, USA). Descriptive statistics are reported as mean, median values and range. Linear Regression Analysis, Student’s t-test were used when appropriate. Non-continuous variables were also compared by Chi-Square test. Variables of survivors and dead patients were analyzed with one-way ANOVA. A P value of <0.05 was considered statistically significant.

Results

The incidence of DP was 4.9% after cardiac surgery. The incidence of DP after open and closed heart surgery were 5.4 and 3.3%, respectively. The incidence of DP which required plication were 1.2 and 1.8% for open and closed heart surgery, respectively. The frequencies of some procedures and the incidence of DP occurring after these specific procedures are summarized in Table 1.

Table 1 The incidence of total diaphragmatic paralysis (DP) cases and DP cases requiring plication in most common operations
Full size table

The overall incidence of DP was higher in correction of tetralogy of Fallot (14.8%, 48/324 patients), Blaloc–Taussig (B–T) shunt (5.7%, 17/294 patients), VSD closure with pulmonary artery patch plasty (7.4%, 17/228 patients) and arteriel switch operation (5%, 8/160 patients) (Table 2).

Table 2 The incidence of overall diaphragmatic paralysis (DP) cases (whether plicated or not)
Full size table

When we searched the incidence of DP, which require plication, we noticed that the distribution of the operative procedure show some differences. Arteriel switch operations were with the highest incidence of DP, which require placation (100%, 8/8 patients). Patients who underwent B–T shunt (58.8%, 10/17 patients) and correction of tetralogy of Fallot (10.4%, 5/48 patients) were also with higher incidence of DP, which require plication (Table 2).

The median age of all patients with DP were 48 months. The mean and median ages of patients who were plicated were found to be significantly young when compared with patients who did not undergo plication due to DP (17.8±11.2 versus 65.3±13.2 months, P<0.0001, median 6 months versus 54 months, respectively).

In 22 patients (52.3%), DP was suspected because of recurrent reintubations due to respiratory insufficiency, while 12 patients (28.5%) failed to wean from mechanical ventilatory support; five patients (11.9%) had recurrent pneumonia. In almost all patients, an elevated diaphragm on chest was present, when observed in the X-ray picture. In patients with bilateral DP, decreased respiratory sounds led us to confirm the diagnosis with further tools like US, fluoroscopy and EMG. The diagnosis was confirmed by fluoroscopy in 109 patients (71.7%), fluoroscopy and US in 29 patients (19.1%), fluoroscopy and EMG in 8 patients (5.2%) and USG alone in 6 (3.9%) patients. The paralysed hemidiaphragm was left sided in 83 (54.6%), right sided in 61 (40.1) and bilateral in 8 (5.2%) cases. Both sides were plicated in patients with bilateral DP. Left hemidiaphragm plication was performed in 24/42 patients (57.1%), right in 10/42 (23.8%) and bilateral plication was performed in 8/42 patients (19%).

The median time from cardiac surgery to plication was 12 days. Mean mechanical ventilation time after cardiac surgery till plication was 14.6±9.8 days (3–63 days). Plication was performed 2.2±1.3 days after the diagnosis. Median age of patients who required plication was 6 months (range 15 days–96 months). Mechanical ventilation time for patients without plication was 3 days (range 1–12 days). Two patients underwent tracheostomy (aged 3 and 5 months). These patients also underwent plication after 17 and 27 days from the surgery, respectively. Both of them died.

Eight patients, who underwent plication died (16.7%)and they were all under 1 year of age. The mean age of dead patients was 5.2±1.3 months. Most of the deaths were due to pulmonary complications even after plication. Only one of these patients had been plicated on the 11th day of surgery and he was 2 months of age. None of the dead patients had bilateral DP. The median plication time for patients who died was 17.5 days and for the survivors 10.5 days. Being under 1 year of age, plication after 10 days from the surgery and pneumonia was found to be significant risk factors for mortality (P=0.019, 0.019, 0.04). The characteristics of patients who died are summarized in Table 3.

Table 3 Summary of deaths
Full size table

The incidence of pneumonia, was significant in patients on whom plication was performed after 10 days from the surgery (P=0.016). The mean and median extubation time of the survivors were 4.2±3.6 and 1.5 days, respectively.

The follow-up period was 22.3±14.5 months. The position of the plicated diaphragm was normal on chest X-rays in all plicated survivors within the 1st, 6th and 12th months after discharge. The patients with DP but did not undergo plication were also evaluated in the follow-up period. Elevated hemidiaphragm on chest X-rays were present but they were still asymptomatic. There was only 5 patients with lung infections among the 110 non-plicated patients in the follow-up period. There was no patients with lung infections among the plicated survivors during the 1-year follow-up.

Discussion

Postoperative course after cardiac surgery in children depends on several situations. Respiratory condition is one of the most important factors influencing mortality and morbidity especially in infants. DP after cardiac surgery is an important complication with negative effects on pulmonary situation [1–9]. Reports of the prevalence of DP after cardiac operations in children vary from 0.3 to 12.8% [3, 6, 10–12, 16, 17]. In our study, the incidence of DP after cardiac surgery was 4.9% and surgically treated DP with transthoracic plication was 1.3%, which is comparable with other series.

Newborns and small infants are seriously affected by DP leading to delayed extubation or requirement of early diaphragm plication [3, 6]. Older children usually have the advantage of handling with diaphragm paralysis as they use their intercostal muscles for respiration when compared with infants [6, 9]. Infants have a more horizontally oriented rib cage and their intercostal muscles are weaker. They also have an increased mediastinal mobility causing mediastinal shifting to the contra lateral side on inspiration with DP. On the ipsilateral side, the diaphragm moves paradoxically as it cannot resist negative intrapleural pressure. This results in decreased functional residual capacity, alveolar collapse and atelectasis, which causedyspnea, recurrent reintubations, failure to wean from mechanical ventilation, retained secretions and pneumonia [8–10, 14]. The majority of our patients (64.3%) who underwent diaphragm plication following cardiac surgery were under 1 year of age. Most common indications for plication were failure in weaning from ventilator or requirement of reintubation. This situation was the major cause of lung infections and related complications such as atelectasis and recurrent pneumonia as reported in other series [3, 5].

Being aware of the anatomical course of the phrenic nerve is essential for the cardiac surgeon as DP occurs most frequently due to technical issues during various operations [3, 6, 8, 9, 12]. After some special procedures, DP was reported to be observed more frequent such as; arteriel switch operation, correction of tetralogy of Fallot, Fontan procedure, repair of total anomalous pulmonary venous connection and B–T shunt [3, 6, 13, 14]. These procedures sometimes require harvesting of autologous pericardium and/or wide exposure of the great vessels. Thymus resection for a clear surgical view, and exposure to the cold solutions have also been associated with phrenic injury [3, 6, 8, 10, 11, 14, 17]. The mean age of the patients who were plicated were found to be significantly young when compared with patients who did not undergo plication due to DP. In our series, after arteriel switch operation, correction of tetralogy of Fallot and B–T shunt; phrenic nerve palsy requiring plication was more common. These operations are generally performed in younger children. In this age group, handling DP by using intercostal muscles is very hard and recovery without plication is very rare. All the causative factors were present in our operations including pericardial resection, extensive dissection, application of ice slush and thymus resection.

Thymus resection was performed with cauter and thermal damage to phrenic nerve may be an important factor for DP even with moderate levels of cauter heat. In our Institute, there was an increase in the incidence of DP over the 10-year period. In the last decade, the age of our patient population decreased and we became the referral hospital especially for cardiac surgery in the infant age such as arteriel switch operations, tetralogy of Fallot or in palliative operations (B–T shunt). As our patient profile became younger, we began to face with these problems increasingly and as we get more experience, we became more suspicious about DP.

Diagnosis of DP is easy if it is suspected. Decreased respiratory sounds in auscultation and elevated hemidiaphragm on chest X-rays of patients will lead the clinician to confirm DP with other diagnostic tools such as US, fluoroscopy and EMG [3, 6]. In our opinion, in patients who fail in weaning from mechanical ventilation, DP should be suspected if there is no cardiac reason. After observing an elevated hemidiaphragm on chest X-rays, further diagnostic study should immediately be performed. In patients with bilateral DP, chest X-rays may not be very helpful in diagnosis. Therefore, decreased respiratory sounds in a patient with respiratory distress should lead the clinician to confirm DP with further tools like US, fluoroscopy and EMG. In our clinic, when weaning is observed to be delayed, we now focus on DP especially in infants. Older children usually tolerate but in infancy, we observe that DP may be mortal and it should be diagnosed before lung infection occurs. All our patients with DP who did not undergo plication has survived, as all of them were older. In the follow-up period, they were healthy except with an elevated diaphragm observed in the chest X-ray. It is not a dilemma, because now we may attribute the unexpected prolonged mechanical ventilations, lung infections and deaths in the past to “undiagnosed” diaphragm paralysis especially in the infant age group. It can be speculated that their clinical course might have been better if they were diagnosed and plicated. We believe that spontaneous recovery is very rare due to persistent images of elevated diaphragms in the chest X-rays in non-plicated survivors. However, we also believe that fluoroscopy in the follow-up period might have been very useful for checking the recovery of DP.

Although it is accepted that surgical plication is an effective way of treatment, there is still controversy on its best timing. The advocates of late plication argue that a waiting period of a couple of weeks may result in spontaneous recovery [2, 8, 12–14], meanwhile some authors recommend to plicate as soon as possible [9], and some recommend to plicate depending on the respiratory condition of the patient [3, 10]. In our study, DP requiring plication in patients under 1 year of age and plication later than 10 days after cardiac surgery were associated with higher incidence of pneumonia and mortality. In our opinion, spontaneous recovery is very very rare. Optimal timing of diagnosis and surgical management of DP in children who have undergone cardiac surgery should not exceed 10 days especially under 1 year of age. The median plication time for our patients who died were 17.5 days and for the survivors 10.5 days. Prolonged mechanical ventilation and waiting for spontaneous recovery will cause unsuccessful weaning attempts leading to increased pulmonary infections, intensive care unit (ICU) stay, cost and mortality. We recommend early plication especially in infants who fail to wean from ventilator with acceptable hemodynamics.

This study is a retrospective, single center study and the retrospective design of our study was a limitation.

As a result, DP due to phrenic nerve palsy after cardiac surgery is a serious complication. The first step in diagnosis of DP is to suspect. X-ray, US, fluoroscopy and EMG are reliable tools and should be performed depending on patient’s condition. Optimal timing in diagnosis and surgical management of DP in children who have undergone cardiac surgery should be made within 10 days. Delay in diagnosis is associated with higher rates of lung infections, prolonged ICU stay time and mortality, especially in patients under 1 year of age. Early plication reduces the duration of mechanical ventilation, with its associated reduced morbidity and mortality especially severe lung infections and ICU stay.