Tracheostomy in critically ill liver disease patients with coagulopathy: A retrospective study at a tertiary center.

INTRODUCTION: Critically ill patients with liver disease commonly present to the intensive care unit (ICU) with need for prolonged ventilation, difficult weaning, and refractory coagulopathy. These patients experience both bleeding and thrombotic complications with a precariously balanced state of coagulopathy. The purpose of this study was to assess the bleeding complications of tracheostomy in critically ill patients with liver disease.
METHODS: A retrospective study was conducted in liver ICU of a tertiary teaching institute. Medical records were analyzed to assess postprocedure complication rate among 73 critically ill liver disease patients who had undergone tracheostomy during the period of October 2017 to September 2018.
RESULTS: Ten out of 73 patients (13%) required transfusion of blood products after 12 h of procedure, despite thromboelastography (TEG)-based correction prior to procedure. Of these, 7 patients (9%) underwent surgical tracheostomy (ST) and three patients (4%) underwent percutaneous tracheostomy. Statistically no significant difference in bleeding was seen among the two groups, but a rising trend was seen with the ST group (P = 0.52). None of the patients experienced procedure-related pneumothorax and subcutaneous emphysema, as observed in the chest X-ray.
CONCLUSION: We conclude that coagulopathy should not be deterrence for the performance of tracheostomy in critically ill patients with liver disease. Adequate clotting support guided by the global tests of coagulation, such as TEG, ensures lesser incidence of bleeding. Copyright:

INTRODUCTION

Tracheostomy is one of the commonly performed procedures in critically ill patients requiring prolonged mechanical ventilation, weaning from assisted mechanical ventilation, airway suction, and airway protection.[1] Percutaneous tracheostomy (PCT) is the popular choice currently due to fewer airway manipulations, low risk of bleeding, and lower incidence of hypoxia and hypercarbia.[1] It also avoids the risk of transport of critically ill patients to the operating room. Surgical tracheostomy (ST) is reserved for cases with difficult anatomy, emergency situations, or when the percutaneous technique fails.[1]

Critically ill patients with liver disease commonly present to the intensive care unit (ICU) with need for prolonged mechanical ventilation and refractory coagulopathy. Liver disease causes changes in both the pro-hemostatic and anti-hemostatic pathways, resulting in a net balance.[2] Therefore, it has been postulated that routine correction of coagulopathy is not required before invasive procedures in patients with end-stage liver disease.[3] However, this balance is precarious and unstable compared to healthy individuals.[3] This explains the potential of bleeding and thrombotic complications in patients with liver disease.

The objective of this retrospective study was to study the incidence of early complications of bleeding, pneumothorax, and subcutaneous emphysema for up to 12 h after tracheostomy in patients with liver disease. We also aimed to compare the complication rate between ST and PCT in patients with liver disease.

METHODS

Study design

After institutional ethics committee approval, retrospective data were collected for adult patients with liver disease admitted in the ICU, who had undergone tracheostomy during the period of October 2017 to September 2018 at our institute. Patients with age <18 years and those undergone tracheostomy in an outside institute were excluded. Data collection was done by going through electronic and scanned records in the medical record department. Demographic characters of age, sex, and etiology of liver disease were noted. Preprocedure international normalized ratio (INR), platelet count, thromboelastography (TEG)-based coagulation index (CI), and blood products transfused were recorded. Severity of illness was assessed with Sequential Organ Failure Score (SOFA), Child–Turcotte–Pugh score (CTP), and Model for End-Stage Liver Disease Sodium (MELD Na) score. Postprocedure complications of bleeding, pneumothorax, subcutaneous emphysema, and blood products transfused were recorded for up to 12 h after the procedure.

Setting

The study was conducted in a 20-bed ICU dedicated to critically ill patients with liver disease. PCT was performed by a single experienced operator. ST was performed in case of short neck, obesity, inability of neck extension, and unavailability of the operator. Decision for tracheostomy was taken by the clinician in charge of ICU. ST was performed by any one of the five surgeons doing the procedure at our institute.

Definitions

ST was defined as the placement of a tracheostomy cannula under direct vision after dissection of pretracheal tissues and incision of tracheal wall in the operation theater. PCT was defined as the blunt dissection of pretracheal tissues followed by serial dilatation of trachea over the guide wire and insertion of tracheal cannula using Seldinger technique at the bedside of the patient. Coagulopathy was defined as a platelet count of <50 × 109 cells/L or INR ≥ 1.5 or TEG derived CI <−3, single criteria or any combination on the day of tracheostomy.[4] Bleeding was defined as continuous stoma bleeding persisting after placement of the tracheostomy tube with hemoglobin drop of at least 1 g% and requiring transfusion of blood products.[5]

Statistical analysis

Mean ± standard deviation was used for normally distributed continuous variables, while median with interquartile range for data if not normally distributed and as percentage for counts. The Chi-square test was used to determine whether abnormal test results were associated with an increased risk of bleeding. Unpaired t-test was used for comparing parametric data between groups and Mann–Whitney rank sum test for nonparametric data. Chi-square analysis and Fisher's exact test were applied as appropriate for comparing proportions between groups. Receiver operating characteristics analysis was used for determining area under the curve and cutoff values. Sensitivity, specificity of clotting index, and laboratory parameters were calculated. All statistical tools were two tailed, and a significant level P < 0.05 was used using IBM® SPSS Statistics version 22 (IBM Corp., Armonk, USA).

RESULTS

There were 723 admissions to the ICU in the study period. Out of these, 280 patients required mechanical ventilation. Seventy-three patients underwent tracheostomy in the study period. All the 73 patients were coagulopathic as per laboratory parameters, of which 10 patients bled (13%). Out of the 73 patients, 49 were found to be coagulopathic according to TEG. Out of the 49 patients with abnormal TEG, 10 (20%) developed postprocedure bleeding necessitating transfusion [Table 1]. Patients those who had significant bleeding had a higher CTP score, elevated INR, a deranged CI on TEG, and an increased duration of mechanical ventilation compared to those who did not bleed [Table 2].

Table 1

Prevalence of bleeding with abnormal reports

Test	Bleeding with abnormal test, number of cases (%)
Laboratory parameters	10/73 (13)
TEG	10/49 (20)

TEG: Thromboelastography

Table 2

Variables between bleeding and nonbleeding groups

Parameters	Bleeding group (n=10)	Nonbleeding group (n=63)	P
Age (years)	49.8±4.2	47.6±7.7	NS
Days of mechanical ventilation*	10.9±0.73	10.15±0.84	<0.05
SOFA score	11.41±0.64	11.20±0.42	NS
CTP score*	10.9±0.31	10.43±0.59	<0.05
INR*	3.12±0.33	2.45±0.43	<0.05
Platelets**	11.10	11.17	NS
CI*	-3.32±0.08	-2.8±0.54	<0.001

**Log of platelets (P value was found with respect to log), *P<0.05. CTP: Child–Turcotte–Pugh, INR: International normalized ratio, CI: Clotting Index, NS: Not Significant, SOFA: Sequential Organ Failure Score

The AUC for INR as predictor of bleeding was 0.88 [Figure 1], with a cutoff of 2.83, showing a sensitivity of 80% and a specificity of 82.5% [Table 3]. The AUC for the CI as predictor of bleeding was 0.92 [Figure 2], with a cutoff of 3.3, showing a sensitivity of 90% and a specificity of 85.7% [Table 3].

Figure 1

International normalized ratio as a significant predictor of bleeding

Table 3

Variables with cutoff

Variable	Cutoff	Sensitivity (%)	Specificity (%)
INR	2.89	80	89
Coagulation index	−3.3	90	85

INR: International normalized ratio

Figure 2

Clotting index as a significant predictor of bleeding

Out of the 73 patients, 42 underwent ST and 31 patients underwent PCT (Griggs method). There was no missing data in any of the patients.

Those undergoing ST were in the higher age group as compared to the PCT group. This difference was statistically significant [Table 4]. The distribution of patients with diagnosis of acute-on chronic liver failure and chronic liver disease was comparable in both groups. No patient with acute liver failure underwent an ST. The severity of liver disease and overall illness was comparable in both groups as shown by CTP, MELD Na, and SOFA scores. The duration of mechanical ventilation prior to tracheostomy did not differ between the two groups. Severity of disease and duration of mechanical ventilation was comparable among the groups [Table 4]. Platelet count and clotting index were comparable in both groups, but the INR was higher in ST group [Table 5].

Table 4

Comparison between surgical tracheotomy and percutaneous tracheostomy

	All (n=73)	ST (n=42)	PCT (n=31)	P
Age, years	51 (29−59)	51 (37−57)	49 (29−59)	0.005
Etiology, n (%)
ALF*	5 (6.8)	0	5 (16.1)
ACLF*	35 (47.9)	22 (52.4)	13 (41.9)
CLD*	33 (45.2)	20 (47.6)	13 (41.9)
Severity of disease
SOFA score*	11 (10-13)	11 (10-13)	11 (10-13)
CTP score*	10.75 (9-12)	10 (9-11)	11 (10-12)
MELD Na*	24 (22-28)	24 (22-28)	24.5 (22-28)
Duration of mechanical ventilation (days)*	10 (8-12)	10 (8-12)	10 (9-12)

*NS. NS: Not Significant,ST: Surgical tracheostomy, PCT: Percutaneous tracheostomy, ALF: Acute liver failure, ACLF: Acute-on chronic liver failure, CLD: Chronic liver disease, CTP: Child–Turcotte–Pugh, SOFA: Sequential Organ Failure Assessment, MELD Na: Model for End-Stage Liver Disease Sodium

Table 5

Coagulation profile in both the groups

Lab parameters	Overall	ST	PCT	P
Hemoglobin, g/dl	8.3 (7.5–10.5)	8.2 (7.5–10.5)	8.3 (7.6–9.2)	NS
Platelets, k/µl	79 (27–176)	79 (27–176)	80 (40.5–132)	NS
INR	2.42 (1.74–3.32)	2.77 (1.74–3.32)	2.34 (1.74–2.94)	0.001
CI	-3.12 (-3.56–-1.18)	-3.16 (-3.56–-1.18)	-3.1 (-3.31–-1.72)	NS

NS: Not Significant, CI: Coagulation index, INR: International normalized ratio, ST: Surgical tracheostomy, PCT: Percutaneous tracheostomy

Complications

Ten out of the 73 (13%) patients required transfusion of blood products within 12 h of procedure [Table 6]. Of these, seven patients (9%) had undergone ST and three patients (4%) had undergone PCT. Statistically no significant difference in bleeding was seen among the two groups, but a rising trend was seen with ST group of patients. None of the patients had pneumothorax and subcutaneous emphysema as observed in postprocedure chest X-ray.

Table 6

Bleeding complications in both the groups

	All (n=73)	ST (n=42)	PCT (n=31)	P
Bleeding complications, n (%)	10 (13.70)	7 (16.66)	3 (9.67)	0.52

ST: Surgical tracheostomy, PCT: Percutaneous tracheostomy

DISCUSSION

Endothelial dysfunction, bacterial infection, and renal failure are considered to be important contributors to coagulopathy in patients with liver disease.[2] Apart from this, the presence of thrombocytopenia and sepsis in critically ill patients with liver disease is found to increase the risk for bleeding.[6] In our study, all our patients fulfilled the definition of coagulopathy according to laboratory parameters, out of which 13% (10/73) had significant bleeding. Among those who were coagulopathic according to TEG variables, 20% (10/49) had significant bleeding. A number of invasive procedures such as central venous cannulation, thoracentesis, and paracentesis have been studied in patients with liver disease to evaluate the risk of bleeding,[47] The results of these studies have established the superiority of the TEG over conventional laboratory parameters for prediction of bleed. Apart from coagulopathy, local complications can also contribute to significant bleeding in these patients, whose coagulation status is precariously balanced.[7]

We found the severity of the liver disease indicated by the CTP score, INR, clotting index, and duration of mechanical ventilation to be predictors of bleeding. Our findings are similar to those of Pandey et al.,[4] who found the severity of liver disease and INR to be significant predictors of bleed. We also found INR of 2.89 to be a good predictor of bleed. Though the duration of mechanical ventilation was found to be significant, there was vast difference in sample size between the groups (10 and 63), which showed a difference of only 1 day, which we don't think is clinically significant. We also found the CI ≤−3.3 to be a significant predictor of bleed. Majority of studies in liver disease have used TEG to guide blood component transfusion rather than prediction of bleed.[78] Pandey et al.[4] had found the K time of the TEG to be a good predictor of bleeding. We have not evaluated the individual parameters of TEG, which is a drawback of our study.

We also evaluated the patients who underwent surgical and PCT. Seven patients had undergone ST and three patients had undergone PCT. Although the difference was not significant (P = 0.52), a rising trend was observed in the ST group. Auzinger et al. conducted a prospective study on sixty patients with liver disease undergoing percutaneous tracheostomy.[9] Their patient cohort consisted of 40% of patients with acute liver failure, followed by chronic liver disease (36%) and post liver transplant patients (20%). Thirteen out of the sixty patients (21%) experienced bleeding complications. Our study is a retrospective data collection of patients with acute-on chronic liver failure and chronic liver disease. The incidence of bleeding was 4% in patients undergoing PCT and 9% for those undergoing ST. Auzinger et al. used preprocedure transfusion of blood products to target INR <1.5 and platelet count of >70 k/μl. We have transfused blood products with the aim of correction of TEG rather than INR and platelet counts. Hence, our patients had clotting indices close to −3.0 prior to the procedure. TEG has been found to be a better predictor of bleeding in patients with liver disease. This could explain the lower incidence of bleeding we have found in our study.

We have also compared the incidence of bleeding patients undergoing ST and PCT. There has been increased incidence of bleed with ST, however this difference was not found to be significant. Therefore, ST can be offered in liver disease patients in whom PCT may not be possible, after adequate correction of TEG. The overall SOFA score and MELD Na were high, indicating the severity of liver disease as well as illness in this group of patients. Auzinger et al. also reported a higher mortality due to increased severity of illness (median Acute Physiology and Chronic Health Evaluation II score = 19) in these group of patients. Assessment of mortality was not a part of our study design. We found that the higher CTP score was significantly associated with the risk of bleeding, but not the MELD Na or SOFA score. This could be due to INR, which is a component of CTP score, and has been found to be a significant predictor of bleed in our study.

We found INR with cutoff value >2.83 to be a significant predictor of bleed. This result is similar to that of Pandey et al.,[4] who found that an INR of >2.6 was a significant predictor of bleed in patients with liver disease undergoing central venous cannulation. TEG variables have been found to be better predictors of bleed in patients with liver disease.[410] They measure changes in clot tensile strength over time and give information on the dynamics of clot formation (coagulation factor and anticoagulant activity), clot strength (platelets and fibrinogen), and clot stability.[11] In our study, we found that, CI with cutoff value <−3.3 was a significant predictor of bleeding in liver disease patients undergoing tracheostomy.

Limitations

The main drawbacks of our study are its retrospective nature, lack of recording of individual TEG variables, and long-term complications.

CONCLUSION

We conclude that coagulopathy should not be deterrence for the performance of tracheostomy in critically ill patients with liver disease. Adequate clotting support guided by the global tests of coagulation such as TEG ensures lesser incidence of bleeding. A prospective study which studies the long-term complications and outcome of this procedure in this group of patients will be more useful.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Ethical conduct of research

This study was approved by the Institutional Review Board / Ethics Committee. The authors followed applicable EQUATOR Network (http://www.equatornetwork.org/) guidelines during the conduct of this research project.