Clinical and Diagnostic Imaging Findings Predict Anesthetic Complications in Children Presenting with Malignant Mediastinal Masses

Abstract

Background

The presence of a mediastinal mass in a child poses significant anesthesia-related risks including death. To optimize outcome clinicians must be able to predict which patients are at highest risk of anesthetic complications.

Methods

We conducted a retrospective review of 118 pediatric patients who presented with mediastinal masses. We investigated their medical records for clinical symptoms and signs at presentation and reviewed their chest radiographs, computed tomography scans, and echocardiograms and electrocardiograms when available. We then conducted analyses to identify clinical and diagnostic imaging features associated with anesthesia-related complications.

Results

Eleven of 117 (9.4%, 95% CI, 4.1%–14.7%) patients experienced an anesthesia-related complication. Four preoperative features were significantly associated with anesthetic complications: orthopnea (p = 0.033, odds ratio (OR) 5.31, 95% confidence interval (CI), 1.15–24.56), upper body edema (p=0.035, OR 8.00, 95% CI, 1.16–55.07), great vessel compression (p=0.037, OR 5.41, 95% CI, 1.11–26.49), and main- stem bronchus compression (p=0.044, OR 5.11, CI 95%, 1.05–24.92). Although the rate of anesthesia-related complications detected in our cohort was comparable to that found in earlier studies, the events were less severe.

Conclusions

Patients who present with orthopnea, upper body edema, great vessel compression, and main stem bronchus compression are at risk of anesthesia-related complications. The low severity of complications in our series may reflect a combination of factors: use of the least invasive method such as interventional radiology to obtain tissue for diagnosis, completion of a thorough preoperative assessment, and minimal anesthetic intervention.

INTRODUCTION

The presence of a mediastinal mass remains one of the most concerning features in the assessment of a child referred for anesthesia or sedation, because of the potential for life-threatening events under anesthesia (1–11). Controversy persists regarding the optimal diagnostic and therapeutic approach in children with a mediastinal mass, and whether anesthesia should be avoided in patients at especially high risk of respiratory compromise (12). Treatment with radiotherapy or chemotherapy prior to tissue diagnosis has been advocated (2), but this approach may compromise the accuracy of the histopathological diagnosis (13). To proceed in an optimal manner in each case, clinicians must be able to predict which patients are at highest risk of anesthetic complications. In this study, we evaluated specific preoperative clinical and diagnostic imaging features and their association with adverse anesthetic outcomes in pediatric oncology patients presenting with a mediastinal mass.

PATIENTS AND METHODS

Patient demographics

We obtained approval for this retrospective review from the St. Jude Children’s Research Hospital Institutional Review Board. The study population comprised individuals identified in the hospital database to have a mediastinal mass at presentation of cancer diagnosed between October 1985 and May 2000. Patients were excluded if the mass was no longer present at the time of admission or if the initial diagnosis of mediastinal mass at presentation was subsequently found to be incorrect. A clinical research associate reviewed the medical records to identify a number of clinical and diagnostic imaging features. Demographic data collected for each patient included age, sex, race, and primary diagnosis.

Medical records review

Clinical symptoms relevant to airway or cardiovascular compromise identified from the medical records comprised cough, dyspnea at rest, dyspnea on exertion, orthopnea, chest pain, wheezing, cyanosis, and superior vena cava syndrome (edema and venous engorgement of the head, upper extremities, or both). Symptoms were recorded as either present or absent. The interpretation of the chest radiograph made at the time of diagnosis was reviewed to establish the mediastinal mass to chest diameter ratio (MMR). The report of the computed tomography (CT) scan of the chest at the time of diagnosis was reviewed to evaluate the compression of the trachea, carina, and mainstem bronchi and to determine whether there was any cardiac or great vessel involvement by the tumor. The preoperative echocardiogram was assessed for pericardial, cardiac, and great vessel involvement. The records were also examined to identify patients who experienced peri-operative complications when the first two anesthetics were administered.

Statistical analyses

Logistic regression models were used to investigate the association between symptoms or factors and peri-operative complications. Associations among factors were analyzed by using Fisher’s exact test and Wilcoxon rank sum test in SAS software (Version 9.1, SAS Institute, Cary, NC). All p-values were two-sided, and a p-value less than or equal to 0.05 was considered statistically significant. A multivariate analysis was not performed because of the low number of complications.

RESULTS

Study cohort

We identified 125 patients who had a diagnosis of mediastinal mass associated with childhood malignancy at the time of admission to our institution. Seven cases were excluded from the analysis: one patient had the mass resected at another hospital before referral to our institution; one patient had multiple pulmonary metastases but no mediastinal mass; two patients had a history of a mediastinal mass, but no mass was found at the time of presentation; and three patients had no mediastinal mass. We therefore analyzed the demographic data of 118 patients (Table 1). The median age at diagnosis of this cohort was 14.1 years (range, 0.4–36.1 years). One patient did not undergo any procedures requiring anesthesia; thus, data from 117 patients were analyzed for associations between preoperative clinical and diagnostic imaging features and peri-operative anesthetic complications.

Table 1.

Demographic characteristics of pediatric patients with a mediastinal mass at presentation (n=118)

Characteristic	n (%)
Race
Black	18 (15.3)
White	97 (82.2)
Other	3 (2.5)
Sex
Female	45 (38.1)
Male	73 (61.9)
Primary Diagnosis
Acute lymphoblastic leukemia	12 (10.2)
Hodgkin lymphoma	45 (38.1)
Non-Hodgkin lymphoma	37 (31.4)
Solid tumor	24 (20.3)

Complications were associated only with administration of the first anesthetic

Eleven of 117 (9.4%, 95% CI, 4.1% – 14.7%) patients experienced an anesthesia-related complication. Although we examined the effects of as many as two anesthetics for each patient, all complications occurred during administration of the first anesthetic; no patient experienced a second complication. The surgical procedures associated with anesthetic complications included: cervical lymph node biopsy (n=1), mediastinal mass biopsy (n=2), central venous line placement (n=1), and bone marrow aspiration and biopsy (n=6). One patient had multiple procedures (Table 2).

Table 2.

Anesthetic Complications

Case No.	Agea (y)	Sex	Weight (kg)	Primary diagnosis	Procedure	Anesthetic	Complication and notes	Treatment
1	22	F	100	HL	Insertion of central line	midazolam, fentanyl	Oxygen desaturation	Unplanned intubation. Naloxone reversal. Extubated after 10 minutes.
2	19	F	55	HL	BM aspiration and biopsy	propofol, fentanyl	Oxygen desaturation to 85%	Recovered without intervention
3	10	M	31	Tcell ALL	BM aspiration	propofol, lidocaine infiltration	PMH of asthma Coughing and wheezing Oxygen desaturation to 95%	Nebulized albuterol
4	5	F	19	Tcell ALL	BM aspiration	propofol, EMLA®	Anesthetized in sitting position Oxygen desaturation to 79%	Suctioned, manually ventilated. Procedure recommenced in left lateral position.
5	6	M	23	NHL	BM aspiration and biopsy	propofol	PMH of asthma Oxygen desaturation to 70%	Manually ventilated
6	17	F	76	RMS	BM aspiration	fentanyl	Large mediastinal mass Intubated in ICU prior to sedation	Apnea requiring change in ventilator settings
7	6	M	27	Tcell ALL	Insertion of central line, BM aspiration, and LP	midazolam, ketamine	Oxygen desaturation to 86% Vomiting	Placed in lateral position, suctioned, face mask oxygen applied
8	12	F	46	Tcell ALL	BM aspiration	fentanyl, propofol, lidocaine infiltration	Oxygen desaturation to 92% shallow respirations and substernal retraction	Recovered without intervention
9	9	M	55	NHL	Percutaneous mediastinal mass biopsy	midazolam, fentanyl, vecuronium	Intubated and difficult to ventilate preoperatively Oxygen desaturation to 79% and difficult to ventilate	On norepinephrine infusion and heliox Manually ventilated
10	13	F	43	NHL	Percutaneous mediastinal mass biopsy	fentanyl, propofol, sevoflurane	Coughing and oxygen desaturation during propofol infusion. Anxiety and oxygen desaturation with emergence whilst supine	Changed to inhalation anesthesia Resolved with sitting position
11	16	M	85	Tcell ALL	Cervical lymph node biopsy	midazolam, fentanyl, nitrous oxide	Airway obstruction Oxygen desaturation to 50%	Intubated and changed to sitting position; oxygenation 90%–100% Extubated after reversal with naloxone and flumazenil

^aAge at diagnosis.

Abbreviations: ALL, acute lymphoblastic leukemia; BM, bone marrow; HL, Hodgkin lymphoma; LP, lumbar puncture; NHL, Non-Hodgkin lymphoma; PICC, peripherally inserted central catheter; PMH, past medical history; RMS, rhabdomyosarcoma

Complications were not associated with any specific anesthetic technique

Anesthetic technique varied, depending on the surgical procedure and perceived severity of the patient’s cardio-respiratory status. Complications occurred under various types of anesthesia: four events occurred under intravenous sedation (combinations of midazolam, fentanyl, and ketamine), six under propofol-based total intravenous anesthesia, and one under combined intravenous sedation and nitrous oxide anesthesia.

Description of anesthesia-related complications

Eleven patients had respiratory complications. Seven patients had minor or transient respiratory problems that responded to airway suctioning, manual ventilation via mask, or change in position to sitting or lateral decubitus. Two patients who were already intubated in the intensive care unit before surgery required an unexpected change in ventilatory strategy during anesthesia. Two patients required unplanned short-term intubation and opioid reversal to treat respiratory compromise.

Complications were associated with preoperative diagnostic imaging and clinical features

The relationship of preoperative symptoms and diagnostic imaging findings to the likelihood of anesthetic complications is shown in Table 3. Four preoperative features were significantly associated with anesthetic complications: orthopnea (p = 0.033), upper body edema (p=0.035), great vessel compression (p=0.037), and main-stem bronchus compression (p=0.044). Two other diagnostic imaging features demonstrated a trend toward a significant relation to anesthetic complications: pleural effusion (p = 0.060) and tracheal compression (p = 0.061). Patients who had no complications had a smaller median MMR (0.37) than did those with anesthetic complications (0.45), although the difference between these values was not significant (p=0.35). The median age of children who had anesthetic complications (10.3 years) was less than that of children who had uneventful anesthesia (14.3 years), although this did not reach statistical significance (p=0.51). We found a relationship between the risk of anesthetic complications and the primary diagnosis of T cell acute lymphoblastic leukemia (p=0.002).

Table 3.

Relations between clinical features and perioperative anesthetic complications in pediatric patients with a mediastinal mass

Clinical Feature (n)a	Proportion of feature (%)	Odds ratio (95% CI) b	P Value
	Clinical presentation
Orthopnea
Negative	107 (91.5)
Positive	10 (8.5)	5.31 (1.15 – 24.56)	0.033
Upper body edema (109)
Negative	104 (95.4)
Positive	5 (4.6)	8.00 (1.16 – 55.07)	0.035
Dyspnea at rest
Negative	94 (80.3)
Positive	23 (19.7)	2.62 (0.70 – 9.84)	0.155
Wheezing
Negative	108 (92.3)
Positive	9 (7.7)	3.14 (0.57 – 17.43)	0.190
Cyanosis
Negative	114 (97.4)
Positive	3 (2.6)	5.20 (0.43 – 62.52)	0.194
Dyspnea on exertion
Negative	89 (76.1)
Positive	28 (23.9)	1.95 (0.53 – 7.24)	0.32
Cough
Negative	78 (66.7)
Positive	39 (33.3)	1.76 (0.50 – 6.19)	0.38
Chest pain
Negative	97 (82.9)
Positive	20 (17.1)	1.09 (0.22 – 5.46)	0.92
	Diagnostic imaging findings
Great vessels compression (93)
Negative	72 (77.4)
Positive	21 (22.6)	5.41 (1.11 – 26.49)	0.037
Mainstem bronchus compression (94)
Negative	72 (76.6)
Positive	22 (23.4)	5.11 (1.05 – 24.92)	0.044
Pleural effusion (95)
Negative	71 (74.7)
Positive	24 (25.3)	4.53 (0.94 – 21.96)	0.060
Tracheal compression (95)
Negative	61 (64.2)
Positive	34 (35.8)	5.09 (0.93 – 27.81)	0.061
Venous engorgement (110)
Negative	107 (97.3)
Positive	3 (2.7)	5.44 (0.45 – 66.04)	0.183
Echocardiogram abnormalities (50)
Negative	39 (78.0)
Positive	11 (22.0)	4.11 (0.51 – 33.27)	0.185
ECG abnormalities (61)
Negative	40 (65.6)
Positive	21 (34.4)	2.90 (0.58 – 14.42)	0.193
Cardiac involvement (93)
Negative	81 (87.1)
Positive	12 (12.9)	1.14 (0.12 – 10.35)	0.91
Carina compression (92)
Negative	76 (82.6)
Positive	16 (17.4)	0.95 (0.10 – 8.70)	0.96

^aUnless otherwise indicated, n = 117.

^bPositive symptom vs. negative symptom; odds of having anesthetic complications

DISCUSSION

This study is the largest single-institution retrospective study to date of anesthesia-related complications in children who have malignant mediastinal masses at presentation, and we found anesthesia-related complications in 11 of 117 patients (9.4%) who underwent anesthesia. Complications occurred with different categories of anesthesia, which, as in other studies (14–19), were not analyzed separately. Most complications were minor, and none were life-threatening. This study is unlike others (14–16,18) in that all patients, including those at risk, received some anesthetic intervention.

Although comparison of peri-operative event rates is limited by variable study methods and definitions, our findings concur with those of four previous retrospective pediatric studies (14,15,18,20). The most recent study found a 15% rate of serious complications of general anesthesia in children with mediastinal tumors (20). In 45 children with lymphoma who underwent general anesthesia or local anesthesia with sedation, 11% experienced intraoperative complications and 6.7% had postoperative complications (18). In another study of 44 children with mediastinal masses who underwent general anesthesia, 18% experienced major peri-operative cardio-respiratory events (15). Investigators in a study excluding patients whose tracheal cross-sectional area was less than half the predicted value found no complications in 42 patients with mediastinal masses (14).

Several adult studies or studies in which data were combined from adult and pediatric patients with mediastinal masses found complication rates from 5% to 18% (17,19,21–23). By contrast, among 177 patients who underwent anesthesia for radiation therapy for cancer in our center between 2004 and 2006, only 49 (1.3%) of 3850 anesthetic sessions were complicated by mostly minor complications (manuscript under preparation). This contrasting finding underscores the importance of judicious approach in sedating patients with a mediastinal mass. Below, we discuss the individual clinical factors and diagnostic imaging findings that may be associated with the risk of anesthesia-related complications in pediatric patients who present with mediastinal masses.

1. Age at diagnosis

In our study, children who had anesthetic complications were younger but not significantly. Other authors have reported that adults with a mediastinal mass are at lower risk of intraoperative complications than children, but a high index of suspicion for the occurrence of early postoperative life-threatening respiratory complications is recommended (17). The only reported direct comparison of the incidence of complications in adults and children found that although adults and children experienced the same overall morbidity and mortality, all deaths of pediatric patients were caused by intraoperative airway complications (22).

2. Primary diagnosis

We found a significant relationship between the risk of anesthetic complications and the primary diagnosis of T cell acute lymphoblastic leukemia (p = 0.002). A recent adult study found no correlation between diagnosis and anesthetic risk (17), while an earlier study in children with mediastinal lymphoma, revealed that only those with non-Hodgkin lymphoma had intraoperative complications (18).

3. Clinical presentation

Orthopnea and upper-body edema were associated with anesthetic complications in our study. In the most recent comparable study of adults, peri-operative complications were predicted by cardio-respiratory symptoms (17).

In two pediatric studies, all patients who experienced intra-operative complications had symptoms preoperatively (18,20). One study found orthopnea to be a particularly good predictor of anesthetic complications (18). The other noted that a combination of three or more respiratory symptoms and signs had a positive predictive value of 50% and a negative predictive value of 100% (20).

Anesthesia-related death and life-threatening airway obstruction have been reported in the absence of preoperative respiratory symptoms (4,6,9,10,24); however, in those cases, either no preoperative CT studies were done (4,9) or imaging demonstrated airway obstruction or a large mass (6,10,24). One report described 20 children who had neither respiratory symptoms nor imaging evidence of tracheo-bronchial compression; all underwent general anesthesia without problems (16).

4. Preoperative superior vena cava syndrome

In our study, upper body edema, a feature of superior vena cava (SVC) syndrome, was associated with anesthesia-related complications. Other reported case series on SVC syndrome have not systematically reviewed anesthetic outcomes (25–27). One such report from our institution described no anesthetic morbidity or mortality in 24 children with SVC obstruction and a mediastinal mass when tissue specimens were obtained by the least invasive feasible method (26).

5. Tracheal cross-sectional area

Reduction of the tracheal cross-sectional area is well known to be an ominous sign. We found an association between mainstem bronchus compression and the risk of anesthetic complications, although we did not quantify the extent of compression as others have done (14,16–18). The association between tracheal compression and anesthetic complications showed a trend toward significance. In a recent study of adults with mediastinal masses, a tracheal cross-sectional area less than half that predicted was associated with postoperative but not intraoperative respiratory complications (17). Similarly, tracheal compression was the strongest predictor of complications of general anesthesia in the most recent pediatric series (20). Other authors have commented on the relationship between tracheal compression and anesthesia complications but have not offered a statistical analysis (14,16,18).

6. Mediastinal m ass ratio

The MMR is used to define small (MMR <0.3), medium MMR (0.31–0.44), and large (MMR >0.45) mediastinal masses (18). In our study we found that patients who had complications had a larger median MMR than those who did not, although this difference was not statistically significant. Three other studies found a relationship between large MMR and peri-operative complications. In one case series of 51 children, all patients who experienced intraoperative anesthetic complications had large tumors (mean MMR, 0.56) (18). In another study of adults and children, the risk of airway obstruction was 2% in patients with a small (≤ 0.31) MMR and 33% in those with a large (≥0 .45) MMR (21). In a retrospective study of 97 adults and children, the rate of peri-operative complications was 19% overall but was 47% in patients with large mediastinal masses (23).

7. Peri-operative involvement of major blood vessels or heart

We found that CT evidence of great vessel compression, but not of cardiac involvement, were significantly predictive of anesthetic complications. Only 50 and 61 of our patients received preoperative echocardiography or electrocardiography, respectively, and neither test was significantly predictive. However, another study found that SVC syndrome accompanied by vascular compression, left atrial compression, and pericardial effusion was significantly associated with general anesthesia complications in pediatric patients (20). One recent adult study found by multivariate logistic regression analysis that pericardial effusion was predictive of intraoperative complications (17). A number of case reports have also raised awareness of the significant risks of anesthesia in patients with cardiac encasement or great-vessel obstruction (3,8,28,29).

8. Pleural effusion

Pleural effusion showed a trend toward statistical significance in our study. Consistent with this, in combination with three or more respiratory symptoms and signs, pleural effusion has been reported to be associated with anesthetic complications (20).

9. Preoperative radiotherapy and chemotherapy

None of the patients in our series received radiotherapy or empiric chemotherapy before anesthesia, however we do not discount the possibility that some patients will be optimally managed using pre-diagnostic radiotherapy. Although pre-biopsy radiotherapy and even chemotherapy have been advocated by some investigators to reduce the size of the mass and therefore the risks of anesthesia (1–3,8,12,21,30), others have cautioned that these interventions distort tissue, thereby compromising the accuracy of diagnosis and curative treatment (13,15).

10. Anesthetic technique

We did not analyze the risk of complications according to anesthesia method. Complications occurred with various types of anesthesia, including intravenous sedation and total intravenous anesthesia based on propofol. Recommendations regarding anesthetic technique in patients with mediastinal masses include maintenance of spontaneous ventilation, avoidance of muscle relaxants, and the immediate availability of various airway manipulation techniques (reinforced tracheal tubes, mainstem bronchus intubation, rigid bronchoscopy, use of heliox, and position adjustment) (2,15,31). In extreme cases, the use of cardiopulmonary bypass has been recommended (32).

Our current practice reflects the following principles: 1) Preoperative evaluation involving multidisciplinary consultation between the oncologist, anesthesiologist, interventional radiologist, surgeon and radiation therapist; 2) Careful choice of the least invasive method available to obtain a diagnostic sample, allowing minimal sedation or anesthesia; 3) Maintenance of spontaneous respiratory effort is critical to avoid the risk of airway collapse in conjunction with positive pressure ventilation and muscle relaxation; 4) Immediate availability of personnel and equipment for emergency airway management including rigid bronchoscopy and tracheostomy, and consideration of cardiopulmonary bypass. These concepts are emphasized in the intervention algorithm (Figure 1).

Fig. 1.

Fig. 1 Intervention Algorithm for Diagnostic Procedures under Anesthesia in Patients with Mediastinal Mass at Presentation

Abbreviations: CXR-chest radiograph; SVCS-superior vena cava syndrome; CT-computer tomography

In conclusion, orthopnea, upper-body edema, main-stem bronchus compression, and great vessel compression predicted anesthetic complications in pediatric patients who presented with mediastinal masses. The overall rate of complications was comparable to previous reports, but the events were less severe. We attribute this finding to a combination of factors, including: (1) use of the least invasive method of obtaining tissue for diagnosis; (2) availability of the technical means to obtain tissue by interventional radiology rather than major surgery; and (3) application of good anesthesia practices, especially a thorough preoperative assessment and the use of minimal anesthetic intervention. These strategies are recommended to minimize the incidence and severity of anesthetic complications in patients with mediastinal masses.