Late Complications After Percutaneous Tracheostomy and Oral Intubation: Evaluation of 1,628 Procedures

The Laryngoscope VC 2015 The American Laryngological, Rhinological and Otological Society, Inc. Late Complications After Percutaneous Tracheostomy and Oral Intubation: Evaluation of 1,628 Procedures Benjamin Storm, MD; Knut Dybwik, RN, PhD; Erik Waage Nielsen, MD, PhD Objectives/Hypothesis: In large international studies, upper airway related stenosis, granulomas, malacias, and laryngeal nerve palsies following percutaneous tracheostomy have an estimated incidence of 6% to 31%. The incidence following prolonged oral intubation is estimated to be 10% to 22%. The purpose of this study was to assess the incidence of late complications in our unit. Study Design: Retrospective search of a single-unit intensive care patient population. Methods: Patient records for a defined period were searched using a predefined search string, identifying those who received invasive mechanical ventilation and split in subgroups by orotracheal tube or tracheostomy tube. This search was cross-linked with the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes associated with recognized complications (J38.0, J38.3,J38.6, J38.7, J39.8, J39.9, J95.0, J95.5, J95.8, J95.9, J99, R04.8,S27.5). Results: During the period January 1, 1997 to December 31, 2013, 32,852 patients were admitted to the intensive care unit. Of these, 1,620 patients received invasive mechanical ventilation. Out of this group, 519 had a tracheostomy and 1,109 were orally intubated. Four tracheostomized and zero orotracheally intubated patients had ICD-10 codes related to complications. From the patient records it became clear that three of four patients with tracheostomy had airway symptoms before being tracheostomized, and the fourth patient had her tracheostomy following a postintubation airway stenosis. Conclusions: Spanning a 17-year period, our study did not show any long-term symptomatic upper airway complications following tracheostomy and only one following orotracheal intubation. This contrasts the internationally estimated incidence. Key Words: Tracheostomy, intubation, late complications. Level of Evidence: 4 Laryngoscope, 126:1077 1082, 2016 INTRODUCTION Since the development and refinement of percutaneous dilatational tracheostomy in the mid-1980s, this method has gained widespread use in the intensive care arena. An estimated 15,000 tracheostomies, of which the majority are percutaneous dilatational tracheostomies, are conducted annually in the United Kingdom. 1 The annual number of tracheostomies in Norway is not known, but an estimated 5% to 35% of mechanically ventilated patients in Norwegian intensive care units (ICUs) receive a percutaneous dilatational tracheostomy. 2 With the increasing popularity of early percutaneous tracheostomies, several studies have raised concerns about potential long-term complications of tracheostomy. From the Department of Anesthesiology and Intensive Care (B.S., K.D., E.W.N.), Nordland Hospital, Bodø, Norway; Institute of Clinical Medicine, University of Tromsø (E.W.N.), Tromsø, Norway; and the Faculty of Professional Studies, University of Nordland (K.D., E.W.N.), Nordland, Norway Editor s Note: This Manuscript was accepted for publication October 5, 2015. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Benjamin Storm, MD, AKUM, Nordlandssykehuset Bodø, PO Box 1480, 8092 Bodø, Norway. E-mail: benjamin. storm@nlsh.no DOI: 10.1002/lary.25766 Tracheomalacia, tracheal necrosis, trachea-esophageal and trachea-arterial fistulas have been described. The two most common long-term complications to percutaneous tracheostomy are granulomas and tracheal stenosis. The incidence of such long-term complications is reported to be around 6% in patients receiving a tracheostomy and 10% to 22% in patients with a prolonged orotracheal intubation, 1 the most common complication being necrosis due to excessive cuff pressure. 3 Most of these cases of long-term complications are probably asymptomatic and hence of minor clinical importance. In one study, bronchoscopy of a series of tracheostomized patients before decannulation showed granulomatous tissue superior to the stoma in 31% of all cases, most of which where asymptomatic. 4 A study by Norwood et al. examining patients up to 80 months (median 21 months) after percutaneous tracheostomy, found 8.3% of the patients had a tracheal stenosis of 26% to 50%. 5 In Norway, healthcare institutions are reimbursed according to activity. Consequently, extensive registers of patient diagnoses and procedures exist in the electronic patient records (EPRs). Searching these codes provides a quick and powerful tool for insight into patient treatment. Our unit is a combined postanesthesia care unit (PACU) and a mixed surgical/medical general adult and pediatric level 2 ICU at a regional hospital. We care for patients from all medical and surgical specialties, 1077

excluding neurosurgical, cardiothoracic, and neonatal patients. We have six ICU beds and treat 275 patients per year. On average, 74 patients per year are mechanically ventilated for more than 5 days (unpublished data). We also care for roughly 4,000 PACU patients per year. The dilatational tracheostomy procedure was introduced at our institution in 1997. All procedures are conducted by an anesthesia staff specialist, a registrar, and one specifically trained clinical nurse consultant (CNC). The procedure is always monitored with bronchoscopy. To ensure proper tracheostomy care, all patients are followed by the same CNC throughout their hospital stay. Since its introduction, the procedure has been increasingly used and there has been a trend toward earlier insertion (unpublished data). We only use silicone tracheostomy tubes with high-pressure cuffs (Bivona TTS; Smiths Medical, Dublin, OH), typically a size 8.0 inner diameter (ID) for men and a size 7.0 ID for women initially. When weaning is commenced, the tube is exchanged to a smaller size to allow better airflow around the tube. For oral intubation, we use only stiff polyvinyl chloride (PVC) tubes with low pressure cuffs (Portex Blue Line Soft Seal Cuff; Smiths Medical). Size 8.0 ID is typically chosen for men and size 7.0 ID for women. Cuff pressure is only monitored occasionally by palpation of the inflation balloon and by listening for leakage. To aid the clinician in deciding on early tracheostomy versus prolonged intubation in patients with expected long-term need for ventilator support, the aim of this retrospective descriptive study of our intensive care population was to assess the incidence of late tracheostomy complications and to compare this to the complications arising from prolonged orotracheal intubation in the same population. MATERIALS AND METHODS A retrospective search and extraction of data from the EPRs using predefined search criteria (Fig. 1) was performed using Qlikview (Qliktech International AB, Lund, Sweden) and/ or the EPR software DIPS (Distributed Information and Patient Data System in Hospitals version 7.3.2.6; DIPS ASA, Bodø, Norway). Patients were identified based on Norwegian Classification of Medical Procedures (NCMP) codes for mechanical ventilation (TG601 and GXAV01), and Norwegian Classification of Surgical Procedures (NCSP) codes for tracheostomy (GBB00 and GBB03) and the World Health Organization s International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes for tracheostomy or intubation related complications, as stated in the literature (Table I). 1,3 5 The period January 1, 1997 to December 31, 2013 was searched. To obtain good descriptive baseline data, the search in the hospital EPRs was cross-matched with a search in our departmental quality database containing detailed information about all tracheostomized patients. This database starts on January 1, 1998; hence, some descriptive data for the year 1997 were not obtainable. Descriptive data for the orally intubated group were obtained through the hospital EPRs. Only data from after January 1, 2002 were extractable. Hence, baseline data for the orally intubated group only covered patients from the period January 1, 2002 to December 31, 2013. Extraction of all-time mortality was not possible. There are some minor 1078 inconsistencies in the data extracted from the EPRs using Qlikview or DIPS. Hence, the patient demographic and descriptive data presented in Tables II and III must be interpreted cautiously. Of note, the data in the two groups do not cover the exact same timeframe. However, it is our belief that the demographics had not changed significantly during the period 1997 to 2002. Data describing ventilator time before tracheostomy were only registered for the years 2003, 2007, 2011, 2012, and 2013. The study was approved by the local data protection officer. Only data relevant to the study have been extracted in anonymous form. Patient Selection EPRs were searched using Qlikview according to the search strategy described in Figure 1. All ICU admissions during the period 1997 to 2013 were identified. Using the NCMP codes for mechanical ventilation TG601 or GXAV01 all patients requiring invasive mechanical ventilation were isolated. Applying NCSP codes for tracheostomy GBB00 or GBB03 this group was divided into patients who received a tracheostomy and patients who did not receive a tracheostomy (Fig. 1). In the search results, all patients ere uniquely identified by a Norwegian Patient Reccord (NPR) serial number. Using a combination of Qlikview, Microsoft Excel 2007 (Microsoft Corp., Redmond, WA), and MS Access 2007 (Microsoft Corp.) on an extract from the hospital EPRs and ICU records (unpublished data), the baseline characteristics and outcome of ICU stay of the two groups was extracted (Tables II and III). The EPRs of the patients were searched using Qlikview for the ICD-10 codes associated with late tracheostomy and intubation complications (Table I). ICD-10 codes related to procedural (i.e., early) complications (bleeding, loss of airway, pneumothorax, and death) were not included. This search identified all patients during the period January 1, 1997 and December 31, 2013 who had a tracheostomy or intubation and a diagnosis of a complication that could have been related to this procedure. Results are presented in Table IV. Using the unique NPR identification number, progress notes from all patients who matched the search criteria were extracted from the DIPS EPRs and read to assess if the complication was related to the tracheostomy/intubation procedure. Results are presented in Table V. Our findings were compared to the incidence of complications reported in other studies. 1,5,6 The findings of the study were discussed with the head of our ear-nose-throat (ENT) department, Dr. Rune Klette, who confirmed that the ENT department had not seen any patients with post-tracheostomy or postintubation problems during the study period. RESULTS During the period January 1, 1997 to December 31, 2013, a total of 32,852 patients were admitted to our ICU/PACU. Of these, 1,620 patients received mechanical ventilation. Out of this group, 519 patients received a tracheostomy and 1,101 patients where mechanically ventilated with an orotracheal tube (Fig. 1). Tracheostomized patients were older (65 vs. 60 years), and a larger proportion were admitted as medical emergencies. Their Simplified Acute Physiology Score (SAPS) was slightly lower than for the orally intubated

Fig. 1. Search strategy. ICD-10 5 International Statistical Classification of Diseases and Related Health Problems, 10th Revision; ICU 5 intensive care unit; NCMP 5 Norwegian Classification of Medical Procedures; NLSH 5 Nordland Hospital. 1079

ICD 10 Code TABLE I. ICD-10 Codes Used for Patient Extraction. Condition J38 Diseases in vocal cords and larynx J38.0 Paresis of vocal cords and larynx J38.3 Other conditions of vocal cords and larynx (including granuloma) J38.6 Larynx stenosis J38.7 Other conditions of the larynx J39.8 Other specified conditions of upper airways (including tracheomalasia type III) J39.9 Unspecified condition in upper airways J95.0 Tracheoesophageal fistula after tracheostomy J95.0 Infection of tracheostoma J95.0 Obstruction of tracheostoma J95.5 Subglottic stenosis after medical procedures (including tracheal stenosis/granuloma and trachea-cutaneous fistula) J95.8 Other specified respiratory disturbances after medical procedures J95.9 Unspecified respiratory disturbance after medical procedures J99.1 Respiratory disturbances in other diffuse connective tissue disorders (including granuloma) R04.8 Bleeding from specified localizations in the airways S27.5 Injury of thoracic trachea ICD-10 5 International Statistical Classification of Diseases and Related Health Problems, 10th Revision. patients (43 vs. 46) (Table II), and they had a lower ICU mortality (15% vs. 20%) and hospital mortality (22% vs. 26%). The tracheostomy was performed a median of 3 days after intubation, and the patients were decannulated a median of 11 days after the tracheostomy (Table III) During the 17-year period investigated, no patients in the orally intubated group had an ICD-10 diagnosis of late postintubation complication. Four patients in the tracheostomized group had an airway-related ICD-10 diagnosis (Table IV). One patient was tracheostomized because of an intubation trauma. This patient was therefore reallocated to the endotracheal intubation complication group. The other three patients had an established diagnosis of airway pathology upon admission to the ICU, and received their tracheostomy as a consequence of this (Table V). They were excluded from further analysis. As a final result, no patients who received a tracheostomy had any diagnosis of airway complication related to this procedure during the study period. The one patient who was reallocated from the tracheostomy to the orotracheal group received a tracheostomy (and subsequent laser surgery to the trachea) due to a postintubation tracheal stenosis. This complication occurred approximately 2.5 months after a traumatic oral intubation and preceded the tracheostomy. This is regarded as a late complication of oral intubation. None (0/516, 0%) of the tracheostomized patients and 0.9& (1/1101) of the orotracheally intubated patients developed procedure-related long-term tracheal complications. TABLE II. Patient Demographics/Baseline Characteristics at Time of Tracheostomy/Intubation. Baseline Characteristic Tracheostomized, n 5 519 Orally Intubated, n 5 1,101 Age, yr, mean (median) 64.9 (67.2) 59.7 (64.4) Male, n (%) 329 (66) 614 (61) SAPS 6 SD 43 6 15.8 46 6 16.6 Missing data, n 0 30 Admission type, n (%) Medical, elective 0 (0) 0 (0) Medical, nonscheduled 274 (55) 553 (55) Surgical, elective 36 (7) 76 (8) Surgical, nonscheduled 140 (28) 352 (35) Missing data 50 (10) 29 (3) Attending specialty at admission to ICU, n (%) Medical (including all 277 (56) 456 (45) subspecialties) Surgical 149 (30) 303 (30) Gynecological/obstetrics 2 (<0,5) 2 (<0,5) Orthopaedic 11 (2) 44 (4) Ear/nose/throat 9 (2) 29 (3) Neurological 50 (10) 134 (13) Pediatric 1 (<0.5) 35 (3) Missing data 0 (0) 0 (0) ICU 5 intensive care unit; SAPS 5 Simplified Acute Physiology Score; SD 5 standard deviation. DISCUSSION Long-term complications of both tracheostomy and intubation are well-described entities. In a study by Norwood et al., 5 patients were examined with bronchoscopy and/or computed tomography of the trachea between 2 TABLE III. Outcome of Intensive Care Stay.* Tracheostomized, n 5 519 Orally Intubated, n 5 1,101 LOS ICU, days, mean (median) 18 (14) 10 (6) LOS hospital, days mean (median) 34 (25) 17 (11) Discharged alive, n (%) 388 (78) 750 (74) Dead in ICU, n (%) 77 (15) 203 (20) Dead in hospital, n (%) 111 (22) 260 (26) Days on mechanical ventilation, 14 (11) 6 (3) mean (median) Days ventilated before tracheostomy, mean (median)* 4 (3) Days tracheostomized, mean (median) 18 (11) Dead during follow-up, n (%) *Data only available for the years 2003, 2007, 2011, 2012, and 2013. Data only available for the period January 1, 1998 to December 31, 2013. ICU 5 intensive care unit; LOS 5 length of stay. 1080

Complication TABLE IV. Occurrence of Late Complications. Tracheostomized, n 5 519 Orally Intubated, n 5 1,101 Vocal cord palsy/damage 1 0 J38.0 & J38.3 Laryngeal stenosis J38.6 & J38.7 2 0 Tracheomalasia type III J39.8 & J39.9 0 0 Tracheal granuloma, stenosis, 0 (1)* 1 (0) or fistula J95.5 and J99.1 Injury of thoracic trachea including 0 0 fistula S27.5 Other respiratory disturbances after intubation J95.8, J95.9 0 0 *One patient reallocated from tracheostomy complication group to orally intubated complication group. and 82 months after tracheostomy. Thirty-one percent of the tracheostomized patients had a stenosis occluding >10% of the tracheal lumen. Of these patients, subjective breathing difficulties were found in 6%. 5 Additionally, in a review by Epstein, long-term complications are stated to occur in up to 65% of patients, of whom 3% to 6% are symptomatic. 4 In light of this, the findings of our study are somewhat surprising. With an estimated 6% incidence of symptomatic complications, we would expect approximately 36 patients to develop symptoms of airway narrowing after tracheostomy. In another study by Rumbak et al., complications after early tracheostomy versus prolonged orotracheal intubation were examined. 7 Our patients are comparable to the patients in both the Rumbak et al. and the Norwood et al. 5 studies with respect to age, sex, and cause of admission, but our patients had longer mechanical ventilation times than the patients in the Rumbak et al. study. Because both Norwood et al. and Rumbak et al. used the Acute Physiology and Chronic Health Evaluation II score and we used the SAPS II, we cannot compare the severity of disease directly. ICU mortality and hospital mortality were substantially lower in our population as compared to the Norwood et al. study (ICU mortality 15% vs. 30%, hospital mortality 22% vs. 40% 6 1%). Our finding of no long-term complications might be related to factors such as the timing of tracheostomy, the technique, the operator experience, the tubes used, the follow-up of tracheostomized patients, and the duration of the tracheostomy. Neither Epstein nor Norwood et al. states what kind of tracheostomy or endotracheal tubes were used in their studies. We only use silicone tracheostomy tubes, which cause less mechanical trauma to the tracheal wall. Our endotracheal tubes are PVC, with a lowpressure, high-volume cuff. Maximum tube size 8 for men and 7 for women are used. To avoid misplacement of the tracheostomy or damage to the tracheal rings and tracheal wall, we use direct bronchoscopy guidance for all dilatational tracheostomies. All tracheostomy procedures are supervised or conducted by an experienced operator (consultant anesthetist or dedicated CNC). These precautions may have resulted in less perioperative tracheal damage and might explain the lower incidence of long-term stenosis and malacia (stenosis is thought to be the result of ischemia and subsequent fibrin formation in damaged tracheal wall, and malacia is thought to be due to cartilage damage during the tracheostomy). 4 In the TrachMan study by Young and coworkers, the outcome differences between early and late tracheostomy was examined. 1 In this study operators spent 20 to 40 minutes on a dilatational tracheostomy. We usually complete the procedure in less than 10 minutes. A long procedure time might reflect serial punctures and other procedural problems, which may, in turn, lead to longterm complications. As mentioned by Epstein, 4 as many as 63% of patients tracheostomized using the Griggs technique (dilatation with forceps) developed >10% tracheal stenosis. Depending on the patient, a stenosis might not become symptomatic before it exceeds 50% of the tracheal lumen. In a series referred by Epstein, symptoms of dyspnea developed <6 weeks after decannulation for more than half the patients and within 2 months for two-thirds of the patients. 4 Until 2011, we also used the Griggs technique at our hospital. Since 2011, we have solely used the Ciglia single-step dilatation technique. Having used the Griggs technique for 14 of the studied 17 years, it is quite surprising that we did not find any patients with any late airway symptoms. This suggests that the development of stenosis is not related to the dilatation technique. The timing of tracheostomy is a controversial subject with no clear international consensus. 8 The timing before insertion varies between 48 hours and 3 weeks. Some studies suggest decreased short-term mortality TABLE V. Patients With a Tracheostomy and Complication Diagnosis. Patient Complication Comment Female, 67 years J38.0 recurrens paresis The paresis was present before surgical tracheostomy. Female, 91 years J38.6 laryngeal stenosis The stenosis was due to radio therapy toward the larynx and was present before surgical tracheostomy. Male, 83 years J38.6 laryngeal stenosis The stenosis was due to a subglottic cancer and was present before the percutaneous tracheostomy. Female, 2 years J95.5 subglottic stenosis after medical procedure The patient received a surgical tracheostomy after a postintubation subglottic trauma. 1081

and morbidity with an early tracheostomy compared to a later. 7,8 The recent TrachMan study could not verify any outcome differences between groups 1 ; however, patients with an early tracheostomy received fewer sedatives and were decannulated earlier. 7 Data from our institution for the years 2003, 2007, 2011, 2012, and 2013 show that our patients received their tracheostomy at day 3.8/2.6 (mean/median), and were decannulated 18.4/11.2 days after the tracheostomy (unpublished data). This is a longer tracheostomy time than that found by Norwood et al. (mean 12.9 days early group and mean 16.1 days late group) and Rumbak et al. (mean 8.4 6 2 days). 5,7 Our patient demographics are roughly equivalent to the Norwood et al. and Rumbak et al. study. Therefore, despite our patients having longer tracheostomy times, our zero complication rate is even more remarkable. In the Norwood et al. study, long-term follow-up occurred at 30 months (range 1 82 months), and the authors state that patients who were contacted were younger at the time of tracheostomy compared to the patients who were not exposed to long-term follow-up. The age of the patients in the follow-up group was 41 years (range 11 83 years) and in the non follow-up group 53 years (range 14 94 years). 6 The tracheostomized patients in our department were significantly older, with an average age of 64.9 years (median 67.2 years) at the time of tracheostomy. This difference means that out patients have died or were dying before they developed symptoms. All ENT hospital service in our health region is provided from Nordlandssykehuset, so it is highly unlikely that any patients with symptoms would be referred to other institutions. It is also possible that the expectation of the patients and the general practitioners differ between our population and the various studies. Both Norwood et al. 5 and Rumbak et al. 7 actively approached and questioned/ examined patients for symptoms and signs of complications. Our study was designed only to identify patients who have contacted the health system with airway symptoms. If we conducted a systematic follow-up interview and/or examination of all tracheostomized (or intubated) patients, it is likely that we would have found a higher incidence of complications. Additionally, we have no information about comorbidities such as diabetes and obesity, which have been linked to complications. If the incidences of these comorbidities are lower in our population, it might help explain the low incidence of complication registered. Despite all the limitations of our study, one would still expect at least some patients presenting with longterm complication to the ENT department. As mentioned, during our study period, no patients presented with or received a diagnosis of complications. This suggests that the incidence of complications is lower in our health region than what is found in the literature. There might also be a substantial negative publication bias in the literature, where studies that find little or no complications tend not to be published. CONCLUSION The incidence of long-term tracheal complications following tracheostomy or intubation in our institution is negligible. On the basis of our study, it appears that both percutaneous dilatational tracheostomies and oral intubations may be safe procedures with few long-term complications. BIBLIOGRAPHY 1. Young D, Harrison DA, Cuthbertson BH, Rowan K; TrachMan Collaborators. Effect of early vs. late tracheostomy placement on survival in patients receiving mechanical ventilation: The TrachMan Randomized Trial. JAMA 2013;309:2121 2129. 2. Sollid SJ, Strand K, Søreide K. Percutaneous dilatational tracheostomy in the ICU: a Norwegian survey focusing on perceived risk and safety attitudes. Eur Jour Anesth 2008;25:925 932. 3. Zias N, Chroneou A, Tabba MK, et al. Post tracheostomy and post intubation tracheal stenosis: report of 31 cases and review of the literature. BMC Pulm Med 2008;8:18. 4. Epstein SK. Late complications of tracheostomy. Respir Care 2005;50:542 549. 5. Norwood S, Vallina VL, Short K, Saigusa M, Fernandez LG, McLarty JW. Incidence of tracheal stenosis and other late complications after percutaneous tracheostomy. Ann Surg 2000;232:233 241. 6. Rosseland L, Laake JH, Stubhaug A. Percutaneous dilatational tracheotomy in intensive care unit with increased bleeding risk or obesity. A prospective analysis of 1000 procedures. Acta Anaesthesiol Scand 2011;55: 835 841. 7. Rumbak M, Newton M, Truncale T, Schwartz SW, Adams JW, Hazard PB. A prospective, randomized, study comparing early percutaneous dilational tracheotomy to prolonged translaryngeal intubation (delayed tracheotomy) in critically ill medical patients.. Crit Care Med 2004;32: 1689 1694. 8. Hyzy RC, Mathur PN, Finlay G. Overview of tracheostomy. Available at: http://www.uptodate.com. Accessed September 2014. 1082