Is Tracheostomy Safe?: Common Indications and Early Complications Among Ugandan Patients In A Pre-Covid-19 Era

Daniel James Nyanzi (  nyanzidaniel4@gmail.com ) Kabale University School of Medicine https://orcid.org/0000-0001-8990-9277 Daniel Atwine Mbarara University of Science and Technology Faculty of Medicine Ronald Kamoga Mbarara University of Science and Technology Faculty of Medicine Caroline Birungi SOAR research foundation Caroline A Nansubuga Makerere University Medical School: Makerere University College of Health Sciences Victoria Nyaiteera Mbarara University of Science and Technology Faculty of Medicine Doreen Nakku Mbarara University of Science and Technology Faculty of Medicine

The many bene ts of tracheostomy have led to a global upward trend in the frequency of the procedure with up-to 250,000 procedures currently performed annually in resource-rich countries (Brenner, Pandian et al. 2020) and indications are being continuously revised. In literature, this rising trend has been attributed to increased access to better intensive care services (Kawale, Keche et al. 2017) and changes in the epidemiology of upper airway obstructive conditions (Adetinuola, Amusa et al. 2011). In the past, obstructive airway disease secondary to acute aero-digestive infections was the most common indication, but in the recent pre-COVID- 19  Like all surgical procedures, tracheostomy carries a risk of adverse events that may increase patient morbidity, prolong hospital stay and add undue strain to an already low resourced healthcare system unclear if this can be generalized to resource limited settings. Furthermore, differences in hospital resources and policies may play a contributory role to variation in patient outcomes and complication rates. For example, in high resource centers, different cadre of staff ranging from surgical residents, anesthesiologists, surgeons and critical care nurses may be trained to perform this complex procedure and care for the patients thereafter. However, in low resource settings like Uganda, it is a reserve for specially trained and in-training medical personnel such as otolaryngologists and oromaxillofacial surgeons while post-procedural care is offered by the nurses and patient attendants. We hypothesize that this may contribute to variation in complication rates for tracheostomy. Other factors that have been documented in literature to be associated with tracheostomy-related complications include; obesity, patient age and tracheostomy type among others ( Therefore, our study aimed at establishing the indications for tracheostomy, the incidence of early tracheostomy-related complications and their associated factors in Uganda.

Methods
We conducted an observational prospective cohort study of adult and child patients who had under-gone tracheostomy at least 2 hours prior to recruitment. A patient was excluded from participation if he or she had: a history of a known bleeding disorder, had no caretaker in hospital, and if they had had the tracheostomy performed from other hospitals besides those involved in this study. In order to achieve the sample size, recruitment was performed at two large university training hospitals in Uganda, that is, Mbarara Regional Referral Hospital (MRRH) in South western Uganda, and Mulago National Referral Hospital (MNRH) in the capital city of Uganda. In both hospitals, the Otolaryngology division is solely tasked with performing tracheotomy procedures.
The sample size of 100 patients was calculated using the formula for estimation of a single population proportion, that is; n = Z 2 *P (1 -P)/r 2 (Kelsey, Whittemore et al. 1996), where: Z = standard normal deviation for two-tailed test based on alpha level (relates to the con dence interval level), assumed at 95% = 1.96; P is the proportion of patients with tracheostomy-related complications, assumed to be 0.42, based on a 42% prevalence of tracheostomy related complications reported in a study in Kenya ( Karuga, Oburra et al. 2012); r is the margin error of estimation that was assumed to be 0.1, that is 10%, plus a 6% addition in sample size to cater for attrition. Ethical clearance was received from the MUST-IREC (MUST-REC 01/11-17) prior to commencement of study activities.
Written informed consent of eligible patients was obtained prior to enrolment into the study. At both sites, patients were consecutively recruited basing on the eligibility criteria two hours post-procedure from different units likely to host tracheotomy patients including the intensive care and high dependence units, emergency units and surgical wards. The 2-hour period was anticipated to be adequate for a patient's post-anesthetic stability and safe transfer from the procedure room to the host ward for post-operative monitoring and care. The decision was further guided by reports that tube obstruction, which is the most anticipated early complication is more likely to occur after 2 hours post-procedure (Nyansikera and Kirui 2013). For patients who were not fully conscious at the time of enrolment, consent was sought from their attendants.
Patients' socio-demographics, indications of tracheostomy, pre-and post-procedural factors including complications were collected at baseline from the patients, their medical records and caretakers using a researcher administered questionnaire. A baseline and day 7 clinical examination was performed so as to establish the complications of tracheostomy. In addition, clinical examination was performed whenever a complication was suspected during the 7 days of follow-up. As a quality control measure, the questionnaire was pretested before its use in the study. The principal investigator also trained the research assistants who were ICU nurses and otolaryngology residents in the study procedures and especially how to recognize the tracheostomy-related complications.
Data was analyzed using STATA version 13.0. Patients' baseline characteristics were described using medians for continuous variables and proportions for categorical variables. Strati ed analysis was conducted to compare the frequency of each indication across hospitals using Pearson chi-square test.
Complications were reported as frequencies and a composite variable of tracheostomy-related complication was generated as a binary variable coded 0 = "No early tracheostomy-related complication", and 1= "early Tracheostomy-related complication". An individual was considered to have an early tracheostomy-related complication if he/she had at-least one of the pre-de ned complications both at enrolment and over the 7 days of study follow up. The overall proportion of patients with at-least one early tracheostomy related complication was calculated as the number of patients who developed at least one complication out of the total number of patients enrolled and expressed as a percentage, while Results A total of 100 patients who had undergone tracheostomy were recruited into the study over a period of one year from MNRH and MRRH. Majority of these were male (70%) and adult patients (84%). Majority of the tracheostomies were performed in ICU/HDU settings (54%), these were emergency tracheostomies in 43% of patients. Peri-operative antibiotics were administered in 60% of patients. Table 1 the incidences of speci c complications were calculated as frequencies of speci c complications out of the total cumulative number of complication events and expressed as percentages. For complications such as tube obstruction which occurred more than once in some patients, only one event was recorded per patient and the total number of reported such events for all patients constituted the total cumulative number of complications.
Using chi-square and binomial regression models in univariate and multivariate analysis, the factors associated with early tracheostomy-related complications during follow-up were established. A signi cance level of 5% was used. Both unadjusted and adjusted risk-ratios were presented in tables with their corresponding 95% Con dence Intervals. Table 1 Sociodemographic and peri-surgical characteristics of patients, overall and by Hospital strati cation airway protection (28%). There were no signi cant disparities in all indications across hospitals except for airway protection which was more commonly encountered in patients at MNRH (40%) as compared MRRH (16%), p=0.008. Table 2 Table 3 shows that upper-airway obstruction was the commonest indication among males (93.7%) while airway protection was the commonest among females (62.5%) at MRRH, and this difference was statistically signi cant, p<0.05. There were also no major disparities in indications of tracheostomy between adult and child patients in MRRH. Overall, at MNRH, no signi cant disparities were noted in indications for tracheostomy across gender and age categories.  Table 4 There was no signi cant difference in the distribution of the common tracheostomy-related complications between MRRH and MNRH except for inadvertent decannulation was commonest at MNRH (26.3%) as compared to 7.9% at MNRH (p=0.037). Factors associated with development of early tracheostomy-related complications Table 5 shows the results of univariate and multivariate analysis for factors associated with tracheostomy related complications.

Discussion
Indications for tracheostomy In this study, we observed that a considerable proportion (69%) of patients had more than one tracheostomy indication. This nding was probably because most (54%) of study participants were very ill ICU and HDU patients who are likely to have multiple challenges for example; reduced consciousness hence unable to protect their airways and clear airway secretions yet at the same time require prolonged assisted intubation, all of which are recognized tracheostomy indications. This study found a comparable distribution of indications at the two study sites probably because of the similarity in patients' sociodemographic characteristics as shown in Table 1, or disease epidemiology and risk for trauma given that they are both urban-based hospitals. We also found pulmonary toilet as the commonest indication which is at variance with what was previously reported in various African studies also had few pediatric patients (16%) and with no patients with aero-digestive infections or airway foreign bodies. Overall, the commonest form of trauma we found was traumatic brain injury rather than aerodigestive trauma reported in other studies. In the same vein, the Indian study had 59% and 21.3% of their patients admitted due to severe organophosphate poisoning and snake bites respectively while in Rwanda, all their patients were recruited from ICU and had history of severe trauma, all of whom were likely to require prolonged periods of intubation for airway support. The present study however did not nd similar indications and the trauma severity was not assessed.
Early tracheostomy-related complications The 53% incidence of early tracheostomy-related complications documented in this study, although in alignment with the global estimated range of 6 to 66%, is higher than the 21.5% reported in Tanzania Incidence of speci c early complications Our study reported tracheostomy tube obstruction as the commonest early complication (52.6%) which agrees with another study in similar settings that reported 80.3% of patients in Kenya (Nyansikera and Kirui 2013). Our nding could be due to the fact that majority (72%) of tubes used had no disposable internal cannula which are known to ease tube care, in addition to absence of heat and moisture exchangers as well as the low sta ng, a challenge that was similarly highlighted by the Nyansikera and Kirui study. This makes it harder to achieve the minimum of 3 times per day tracheostomy tube suctioning as recommended (Morris, Whitmer et al. 2013) to prevent this complication. Life-threatening complications like tracheo-innominate stula, pneumothorax and pneumomediastinum were not reported in this study and no mortality occurred as a result of tracheostomy.
Factors associated with development of early tracheostomy-related complications Bearing in mind tracheostomy tube obstruction as the commonest complication in this study, our nding that anticipated prolonged intubation as an indication increased the risk of early complications 1.8 times compared to pulmonary toilet may be because patients who were operated for this indication were more likely to have had a signi cantly severe illness. Such patients would still be too sick to self-care for their tracheostomy tubes even in the post-procedure period and this may be compounded by their inability to communicate a need to be suctioned or have the tube cleaned or changed in addition to a higher possibility of a poor cough re ex. All these may have increased chances of retention of copious airway secretions hence a higher risk of tube obstruction. Such patients are also more likely to require assisted bed turning and bathing compared to more stable patients and this increases chances of inadvertent tube decannulation, which was in fact the second most common complication (17.1%) in this study.
Our nding that patients younger than 18years had an increased risk of early complications was not surprising since this association had previously been reported by a study in a similar setting that found a higher complication rate among children below 10years as compared to older patients (Gilyoma, Balumuka et al. 2011). This is in agreement with our study in which 10 out of the 16 children enrolled were below 10 years of age. This age group could be more prone to complications because of their anatomical features such as a shorter neck, more pliable laryngeal structures and more prominent subcutaneous fat. Also, pediatric tracheostomy requires procedural modi cations if complications are to be minimized. For example, neck hyperextension may pull mediastinal structures in to the neck thus increasing chances of damage to lung apices and resultant emphysema, pneumothorax and pneumomediastinum. Also, awake pediatric patients are still less likely to communicate the need to be suctioned which increases chances of tube obstruction as compared to adults.
The Bjork ap is an inferiorly based tracheal ap through the 2nd, 3rd, and 4th tracheal rings which is xed to the skin to stabilize the tracheal lumen. Although it has a risk for tracheal stenosis, it has advantages over traditional incisional and excisional window procedures including; reduced risk of false cannulation especially during emergency recannulation following accidental tube dislodgement, early stomal maturation and ease of stomal care by assistants and family members ( Our nding that Bjork ap and vertical tracheal incisions were associated with increased risk of early complications and the fact that tracheostomy tube obstruction was the commonest complication across all tracheal incisions including patients with Bjork ap (7 out of 9) cannot be conclusively explained by this study. There is need to further explore this relationship among patients who receive Bjork ap.
The strength of our study is that it gives information on tracheostomy-related complications in Uganda within a period preceding the COVID-19 pandemic and so could be used as baseline assessment for future studies that will assess the impact of COVID-19 pandemic on tracheostomy outcomes. However, since the study was done in the pre-COVID-19 period, its ndings may need to be interpreted with caution within the current COVID-19 context. Also, the small sample size may have limited its ability to establish true associations between complications and the risk factors identi ed. Furthermore, the small number of pediatric patients in this study may affect the generalizability of our study results to the pediatric age group.

Conclusion And Recommendation
Within a pre-COVID-19 context, pulmonary toilet is the commonest tracheostomy indication at major hospitals in Uganda. The incidence of early tracheostomy complications is high and majorly related to post-procedure tracheostomy tube management. Having anticipated prolonged intubation as an indication for tracheostomy, a Bjork ap or vertical tracheal incisions and being a child increases the risk of complications. Emphasis on surgical training and post-procedural management for both clinicians and nurses should be made. Future studies will be needed for better understanding about the possible impact of the COVID-19 pandemic on the outcomes of tracheostomy in the same context.

Consent for publication:
Not applicable Availability of data and materials: