Rebleeding After Secondary Post‐Tonsillectomy Bleeding: Risk Factors Identified From a Single‐Centre Retrospective Study of 448 Cases

Tonsillectomy is one of the most common elective surgical procedures. Post-tonsillectomy haemorrhage (PTH) can be minor with spontaneous resolution, or more severe with risk to the airway and hemodynamic stability, necessitating immediate surgical intervention. PTH is typically divided into primary or secondary bleeding depending on whether it occurs within 24 h. In children, primary bleeds have a reported rate of 0.2%–2.2%, while secondary bleeds, a reported rate of 0.1%–3.0% [1]. The rate of primary and secondary bleeds are generally higher in the adult population [2]. Typically, patients with secondary PTH are admitted to hospital overnight, mostly owing to the risk of rebleeding [2]. A variety of predictors of PTH have been explored in the literature [2]; however, there is limited research on risk factors for, rates of and management of rebleeding following an initial PTH. Two studies reported rates of rebleeding after PTH to be between 10% and 12% [3, 4]. A retrospective study of 337 paediatric patients with a history of PTH but normal examination in the emergency department, reported an 11% chance of rebleed, with 84% of rebleeds occurring within 24 h of presentation with PTH [4]. This led the authors to advocate for close monitoring for the first 24 h after presentation with PTH. In another study of 67 patients who presented to the emergency department with PTH, but had a normal clinical exam, there was a 3% rate of return to theatre [5]. The authors therefore argued that patients who present to an emergency department with PTH but with a normal exam may not require admission. These two studies, therefore, made directly contradictory recommendations. Given the limited and conflicting research into predictors of rebleeding following PTH, the aim of this study was to identify risk factors for rebleed following secondary PTH with the objective of directing best management in the context of this common complication. Ethical approval for this study was obtained from the Auckland Health Research Ethics Committee (reference number AH2432). The reporting guideline used in the manuscript preparation was STROBE. All patients, including both adults and children, presenting to Waikato Hospital, a tertiary public hospital in Hamilton, New Zealand with PTH from January 2005 to January 2022 were identified. This period was chosen owing to the availability of online and paper notes. Exclusion criteria were primary post-tonsillectomy bleeds, inappropriately coded data in the electronic database, missing information regarding rebleeding or examination findings, all patients with objective evidence of a bleeding tendency and radical tonsillectomy for oncological reasons. The electronic and paper clinical notes of each patient were examined by a medical student researcher, not employed by the Hospital, to minimise bias. Demographic data, namely age, sex, ethnicity and New Zealand Deprivation Score (NZDep score) [6], as well as clinical data, namely body mass index (BMI), number of days post tonsillectomy, estimated PTH bleed volume and examination findings (normal, coagulum or active bleeding) at the time of presentation with PTH, were collected for each patient. NZDep score is a flexible area-based measure of socio-economic deprivation in New Zealand, and frequently serves as the primary social factor examined in studies on population health and social research in New Zealand. It incorporates a range of data including measures of employment, income, crime, housing, health, education and geographical access to give an overall measure of deprivation [6, 7]. Other treatment-related data collected included indication for tonsillectomy, status of the tonsillectomy surgeon, simultaneous adenoidectomy or not, use of tranexamic acid (TXA) and non-steroidal anti-inflammatory drugs (NSAIDs) post-operatively and interventions taken following PTH: tranexamic acid, hydrogen peroxide (H2O2) gargles, silver nitrate (AgNO3) cautery and observation alone. All further episodes of bleeding, ‘rebleeds’, were recorded, regardless of whether this was during or after the re-admission period, including time to rebleed following initial PTH. Given that many of the patients reported on in this study had their primary surgery performed in outside institutions, including local private hospitals, it is not possible to measure the overall number of tonsillectomies performed over the study period and therefore we cannot report the overall rate of PTH. Statistical analysis was performed using SPSS (IBM SPSS Statistics 29). For univariate analysis, Mann–Whitney U-tests and chi-square tests were used to assess impact on risk of further rebleed. An unpaired t-test was used to compare NZDep in those that did and did not return the theatre. Following univariate analysis, a logistic regression model was constructed using variables that either had a statistically significant impact on risk of rebleed or those that were clinically thought to have a clinical impact on risk of rebleed. Statistically significant was defined as a p-value of < 0.05. In total, 635 admissions due to PTH were identified through searching codes in the local electronic database. Of the patients excluded (note some patients were excluded for more than one reason), 127 had a primary tonsillectomy bleed, 71 were coded incorrectly, 52 had missing examination findings data, 50 had missing rebleed data, 12 had altered bleeding tendencies and 3 were patients who underwent radical tonsillectomy. These data are summarised in Figure 1. After exclusions, 448 patients, including 257 female (57.4%), with mean age of 16.8 years (SD: 9.99 years) were included in the analysis with the demographic and clinical data shown in Table 1. New Zealand European/Pakeha made up the majority of patients (68.1%), with New Zealand Māori following next at 21.7%. The majority of patients (68.1%) had a NZDep score of 6–10, as shown in Figure 2. The mean number of days from surgery to PTH was 5.73 days (SD: 2.574). Sixty-nine patients (15.4%) exhibited rebleeding after a PTH had occurred. In the univariate analysis, the only factors that had a statistically significant impact on risk of rebleed were higher NZDep score (18.0% in higher NZDep score group vs. 9.8% in the lower NZDep score group; p = 0.024), and intervention with AgNO3 cautery (24.7% in those who underwent AgNO3 cautery vs. 13.6% in those who did not; p = 0.017). The result of the univariate analysis is shown in Table 1. The use of AgNO3 at presentation with PTH occurred in 9 out of 156 patients (5.8%) with normal examination findings, 46 out of 229 patients (21%) with coagulum and 18 out of 63 patients (29%) with active bleeding, as demonstrated in Figure 3. Of those 18 patients with active bleeding who underwent AgNO3 cautery, 6 (33%) subsequently required a return to theatre. Of the 46 patients with a coagulum on examination who underwent AgNO3 cautery, five subsequently required a return to theatre and of the nine patients with a normal examination who underwent AgNO3 cautery, zero subsequently required a return to theatre. A breakdown of AgNO3 usage by age is also shown in Table 2. Among the 69 patients who experienced a rebleed, the median number of hours to rebleed was 19 h (range 1–240 h), with 21 patients (30.4%) experiencing a rebleed within 6 h and 46 patients (66.7%) experiencing a rebleed within 24 h of their initial PTH, as shown in Figure 4. Of those 69 that re-bled, 23 subsequently returned to the operating theatre for haemostasis. Thirty-six patients who presented with PTH returned to theatre directly for haemostasis and did not exhibit further rebleeding (as illustrated in Figure 1). For patients who returned to theatre, the mean NZ Dep score was 6.56, with a standard deviation of 2.71. For patients who did not return to theatre, the mean NZ Dep score was 6.86, with a standard deviation of 2.63 (p = 0.4171). A logistic regression model was constructed as shown in Table 3. The only variable with a statistically significant impact on risk of rebleed was NZDep score, with an odds ratio of 2.35 (95% confidence interval 1.21–4.57; p = 0.011). There are contradictory studies in the literature regarding how to manage PTH in the context of the risk of rebleeding and the need for hospital admission [4, 5]. This study represents a large case series of patients describing the risk and pattern of rebleeding after a PTH has occurred. It makes several important observations to guide management. We report the findings from 448 patients, with a gender and ethnic distribution consistent with the New Zealand general population [8]. It seems likely that these findings should be generalisable to other Western societies. This is a retrospective study with inherent limitations although it is noted that the number of patients excluded owing to insufficient data wererelatively low. We did also choose to exclude patients who have a bleeding tendency as well as radical tonsillectomy and so the data need to be interpreted in that light. Sixty-nine (15.4%) patients experienced rebleeding after the initial PTH. This number captures all episodes of rebleeding after PTH which, when compared to other studies that merely examined much smaller numbers of repeat hospital presentations, generates additional statistical power [9]. Analysis of our data showed NZDep score and use of AgNO3 cautery as the only factors to be statistically significant predictors of rebleeding. Analysis of NZDep scores and return to theatre rates suggest that there is no statistically significant relationship between a lower/higher NZDep score and a patient's return to theatre rate. The risk of rebleed was 9.8% for patients in the less deprived half, while those in the more deprived half had approximately double that risk at 18.0%. While socio-economic deprivation has been associated with poorer outcomes in many other healthcare studies, this is the first time this association has been noted in the literature for rebleeding after PTH. The ways in which social factors impact on health have long been recognised as being multiple and varied, spanning factors as diverse as diet, housing, education, sanitation and employment as well as access to healthcare [10]. The mechanisms by which deprivation might contribute to this observed effect are therefore numerous. For instance, a retrospective study of 274 patients demonstrated a positive correlation between post-tonsillectomy pain and PTH in adults [11]. While the analgesic regime used is not standardised across our region, it is notable that post-discharge analgesia for these patients would have been at their own expense. It is conceivable that inadequate access to analgesia in those that are more socially deprived is a factor. Although fundamentally this is not a study about risk factors for PTH itself and we have only studied those patients presenting with PTH, thus bringing in a selection bias, it is noted that presentation with PTH was also skewed towards those experiencing greater deprivation. This merits further study. Use of AgNO3 cautery to control bleeding in PTH is predictive of a rebleed. Patients who had AgNO3 cautery had a 24.7% chance of rebleed, while patients who did not, had a rebleed rate of 13.6%. This is consistent with the findings of a retrospective study of 145 paediatric patients, which reported that 34% of patients treated with AgNO3 had immediate or eventual return to the operating room for haemorrhage control, compared to the 11% in the no intervention group [12]. While AgNO3 use is associated with rebleed risk, this does not prove causation and its use may merely be a marker of a higher risk patient. Our data do not necessarily imply that use of AgNO3 is unsafe and it is notable that in the context of active bleeding and cautery with AgNO3, 67% of patients did not subsequently require a return to theatre. We would recommend however that where AgNO3 is used, clinicians should anticipate a higher risk of further problems. It has been routine practice in our institution to admit all cases of PTH for a period of 24 h. We have demonstrated that a significant proportion of rebleeding occurs within 6 h, corresponding to a period of observation in the emergency department setting, as well as a proportion occurring more than 24 h later. We have therefore moved to a practice of selective discharge after a short period of observation in the emergency department, where PTH occurs and the patient is deemed to be low risk. Unfortunately, our data do not allow an evidence-based protocol to clearly guide decision-making in this context and good clinical judgement is still recommended. Where clinicians consider discharging patients with PTH from the emergency department, we recommend considering socio-economic status and whether AgNO3 has been used. M.L. acquisition of data, data analysis and interpretation, drafting and revision of manuscript. A.W. conception, design, interpretation and revision of manuscript. O.A. conception, design, interpretation and revision of manuscript. B.W. data analysis and interpretation of results. K.S. conception, design, interpretation, final review of manuscript and person accountable for all aspects of the work. The authors have nothing to report. This study obtained ethics approval from the Auckland Health Research Ethics Committee (reference number AH2432), and the Clinical Audit Support Unit at Waikato Hospital. The authors declare no conflicts of interest. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/coa.14275. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.