Introduction
Tonsillectomy is a common operation performed largely on children.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 The principal indications are recurrent bacterial tonsillitis, peritonsillar abscess (quinsy) and obstructive sleep apnoea.Reference Stuck, Gotte, Windfuhr, Genzwurker, Schroten and Tenenbaum 2 For these indications, tonsillectomy has been shown to significantly improve patient outcomes, thereby demonstrating its clinical value.Reference Konieczny, Biggs and Caldera 3 In practice, extensive clinical experience is required to determine which individuals should be offered surgery because of difficulties interpreting child and parental complaints. Tonsillectomy is almost entirely performed as an elective procedure.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1
Tonsillectomy is a low-risk procedure.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 Most complications are not serious; they include excessive pain, vomiting, odynophagia and fever. Post-tonsillectomy haemorrhage remains the most serious complication, which can occur at any time in any patient, and has the potential to be life-threatening. A recent Swedish study identified the mortality rate associated with post-tonsillectomy haemorrhage as 1 in 41 000.Reference Ostvoll, Sunnergren, Ericsson, Hemlin, Hultcrantz and Odhagen 4 Post-operative haemorrhage is classified as primary (occurring less than 24 hours after the procedure) or secondary (occurring more than 24 hours after surgery). The causation of primary haemorrhage is widely accepted to be suboptimal surgical technique.
The rate of post-tonsillectomy haemorrhage reported in the literature is highly variable. This variation arises because of differences in: the study patient population, the length of post-operative observation and the definition of ‘post-operative haemorrhage’.Reference Stuck, Gotte, Windfuhr, Genzwurker, Schroten and Tenenbaum 2 , Reference Windfuhr, Verspohl, Chen, Dahm and Werner 5 The UK post-tonsillectomy haemorrhage rate in 2003–2004 was 3.5 per cent (n = 33 921; primary = 0.6 per cent, secondary = 2.9 per cent), with 0.9 per cent of affected patients returning to the operating theatre.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 In 2003–2004, the most commonly used technique in the UK was cold-steel with bipolar haemostasis (35 per cent; overall rate of post-tonsillectomy haemorrhage = 2.7 per cent), followed by bipolar forceps (30 per cent; haemorrhage = 4.6 per cent), cold-steel with ties (13 per cent; haemorrhage = 1.7 per cent), and coblation (5 per cent; haemorrhage = 4.6 per cent).Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1
Several risk factors have been suggested for post-tonsillectomy haemorrhage. The National Prospective Tonsillectomy Audit (2005) identified an increased risk of post-tonsillectomy haemorrhage with increasing patient age, male gender and with single-use instruments.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 Other studies have identified known coagulopathy and increasing operative time as risk factors.Reference Stuck, Gotte, Windfuhr, Genzwurker, Schroten and Tenenbaum 2 , Reference Windfuhr, Verspohl, Chen, Dahm and Werner 5 The tonsillectomy technique used has been suggested as a risk factor by many studies; however, the findings of a recent Cochrane review were inconclusive.Reference Windfuhr, Verspohl, Chen, Dahm and Werner 5 – Reference Mosges, Hellmich, Allekotte, Albrecht and Bohm 7 The National Prospective Tonsillectomy Audit found that ‘hot’ (bipolar or coblation) techniques were associated with a greater risk of haemorrhage than cold-steel with ties, although only the coblation technique was associated with a statistically significant elevated risk of returning to the operating theatre.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 The National Prospective Tonsillectomy Audit found no statistically significant relationship between the grade of a surgeon and risk of haemorrhage.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1
As coagulopathy is a potentially modifiable risk factor for post-tonsillectomy haemorrhage, it is often speculated whether all patients should be investigated. Stuck et al. suggest that to detect coagulation disorders, a focused history is of greater benefit than routine coagulation testing.Reference Stuck, Gotte, Windfuhr, Genzwurker, Schroten and Tenenbaum 2 , Reference Koscielny, Ziemer, Radtke, Schmutzler, Pruss and Sinha 8 Therefore, if there is no suggestion in the history, routine coagulation screening is not indicated.Reference Stuck, Gotte, Windfuhr, Genzwurker, Schroten and Tenenbaum 2
All patients are at risk of unsafe medical practices and care during procedures in hospital. Multiple studies have identified an increased mortality risk for medical and surgical emergency weekend admissions compared with weekday admissions.Reference Aylin, Yunus, Bottle, Majeed and Bell 9 , Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 This has been attributed to reduced staffing and service availability at the weekend, in combination with ‘sicker weekend patients’. Mohammed et al. reported weekend admission to be an independent risk factor for in-patient death; the risk was more pronounced in the elective setting.Reference Mohammed, Sidhu, Rudge and Stevens 11 Further research has identified increased 30-day mortality rates for elective surgery carried out closer to the end of the week and during the weekend itself.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 It is important to note that other studies have found no such effect.Reference Becker 12 – Reference Roberts, Thorne, Akbari, Samuel and Williams 14 In addition to investigating high-risk procedures within general surgery and cardiothoracic surgery, Aylin et al. also analysed mortality for low-risk procedures including tonsillectomy.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 The authors found higher mortality rates for low-risk procedures carried out on a Friday compared with those conducted on a Monday, but there was no significant difference for procedures performed at the weekend.
Our aim was to investigate whether the findings of these studies had any relevance within otolaryngology, using tonsillectomy as the chosen procedure. It is inadequate to use mortality rate as an outcome in low-risk procedures such as tonsillectomy. The occurrence of complications such as post-tonsillectomy haemorrhage may provide a better indicator of altered patient care and is more relevant to everyday practice.
With increasing pressure for National Health Service trusts to meet government waiting-time targets, the number of elective surgical procedures performed at the weekend has increased. 15 There is a danger of having too many cases on these additional surgery schedules, which introduces further time pressure. As some studies have indicated an increased mortality risk in patients who undergo elective surgical procedures at the weekend, we hypothesised that post-tonsillectomy haemorrhage rates would increase for patients undergoing surgery at the weekend.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 The rationale for this is poorer quality of care at the weekend as a result of reduced staffing levels and less senior or experienced staff.
Materials and methods
Routinely collected hospital data for a three-year period, between 1 January 2010 and 31 December 2012, from Lewisham Hospital, London, were analysed. The total numbers of elective tonsillectomies performed were retrospectively extracted from electronic operating theatre records. The day of operation was identified for each procedure. The same group of surgeons operated principally on weekdays, but also on the weekend because of government waiting-time initiatives. The records grouped cold-steel and bipolar tonsillectomy techniques as one, but distinguished coblation tonsillectomy.
The emergency department records were used to identify all patients who presented with a post-tonsillectomy haemorrhage. These patients were cross-referenced with the elective tonsillectomy patient list to ensure the initial procedure had been performed at Lewisham Hospital and not elsewhere. Patients whose original operation had not been performed at Lewisham Hospital were excluded.
Electronic operative records were used to identify which patients had returned to the operating theatre following a post-tonsillectomy haemorrhage and this list was cross-referenced with the collected data. This enabled us to identify patients who had returned to the operating theatre prior to being discharged following the operation.
The data collected included patient characteristics, surgical technique, grade of surgeon, time of bleed following tonsillectomy, and requirement of operative arrest of haemorrhage.
A primary post-tonsillectomy haemorrhage was defined as a bleed that occurred within 24 hours of the operation; it included those patients in whom the haemorrhage started prior to hospital discharge and those who returned to the emergency department within this timeframe. All other post-tonsillectomy bleeds were classified as secondary haemorrhages.
Statistically significant differences between the weekday and weekend haemorrhage rates were calculated using the chi-square test. Demographic data were compared using the chi-square test, for categorical variables, and the t-test, for continuous variables. Differences were considered statistically significant at p < 0.05.
Results
From 1 January 2010 to 31 December 2012, 2349 tonsillectomies were performed at Lewisham Hospital. Over this period, 114 (4.86 per cent) post-tonsillectomy haemorrhages were recorded. Of these, 100 (87.72 per cent) were secondary haemorrhages. There were 14 (12.28 per cent) primary haemorrhages. Fifty-two patients (2.21 per cent) required re-intubation for haemorrhage arrest in operating theatres. There were no mortalities.
The coblation technique was used in 35.70 per cent of cases; the bipolar or cold-steel dissection techniques were used in the remaining cases (64.30 per cent).
Of the tonsillectomies, 94.00 per cent were performed on weekdays (Monday to Friday). The remaining 6.00 per cent were performed on the weekend: 4.29 per cent on Saturday and 1.79 per cent on Sunday.
There was no significant difference between the weekday and weekend surgery groups in terms of primary (p = 0.343) or secondary (p = 0.0855) post-tonsillectomy haemorrhage (Table I).
Weekday versus weekend post-tonsillectomy bleed rates
The weekend surgery patients had different demographics to the weekday surgery patients. Overall, the weekend surgery group were older, predominantly female, underwent fewer coblation technique procedures, and had a greater proportion of surgical procedures performed by a consultant or associate specialist (Table II).
Weekday versus weekend tonsillectomy demographics
*n = 2208; † n = 141. AS = associate specialist
Patient age was significantly higher (p < 0.001) in patients who experienced a post-tonsillectomy bleed (18.76 years) compared to those who did not (12.57 years). There was no significant difference between those who suffered a post-tonsillectomy bleed compared to those who did not in terms of gender, operative technique or grade of surgeon (Table III).
Post-tonsillectomy bleed risk factors
*n = 2235; † n = 114. AS = associate specialist
Discussion
Our primary and secondary post-tonsillectomy haemorrhage rates are in line with those reported in current literature.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 Only two of the surgeons were trained in the coblation technique, which accounts for its overall reduced use. As expected, the vast majority of procedures were performed on a weekday.
The significant difference in demographics between the two groups is important. The difference in age between the weekend and weekday surgery group is most likely a result of patient selection, with the older children or adults being selected for weekend surgery. The cause for the difference in sex is unclear, but is probably incidental given the relatively small number of weekend cases. A reduced number of coblation tonsillectomies were performed on the weekend; this is expected, as only two of the surgeons operated using this technique. These two surgeons performed a large proportion of the weekday tonsillectomies. The weekend tonsillectomies were all performed by senior surgeons, either consultant or associate specialists; training registrars were not invited to participate in weekend surgery as they would require supervision.
The only risk factor for post-tonsillectomy haemorrhage in this study was age: mean patient age was higher in the post-tonsillectomy haemorrhage group (p < 0.001; Table III). This is consistent with previous study findings.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 There was no significant difference in post-tonsillectomy haemorrhage rate as a result of patient gender, operation technique or surgeon grade. Hence, although the weekday and weekend surgery groups were different in terms of demographics, these differences are unlikely to have affected the overall results of the study. Previous studies have found that male gender is associated with an increased risk of post-tonsillectomy haemorrhage; we did not find this to be the case.Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1
Older age was associated with an increased post-tonsillectomy haemorrhage rate. In addition, the patients who underwent tonsillectomy on the weekend tended to be older. Although there was no significant difference between the weekday and weekend tonsillectomy groups, a higher percentage of secondary post-tonsillectomy bleeds occurred at the weekend (7.09 per cent vs 4.08 per cent); this could be attributed partly to the older age of patients who underwent surgery at the weekend.
In this study, there was no statistically significant increase in the number of post-tonsillectomy haemorrhages when the procedure was performed on a weekend compared to a weekday.
A number of studies have investigated the effects of procedure day related to mortality for elective surgery.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 , Reference Mohammed, Sidhu, Rudge and Stevens 11 In 2013, after analysing all English hospitals that undertook elective surgery over three years, Aylin et al. concluded that there appeared to be a higher risk of death for patients who have elective surgery carried out later in the working week and at the weekend.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 This was true principally for high-risk procedures, which included oesophagectomy and coronary artery bypass grafts.
The reason for increased mortality at the weekend is unclear and is suggested to be multifactorial.Reference Aylin, Yunus, Bottle, Majeed and Bell 9 , Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 , Reference Becker 12 In general, fewer staff are present at the weekend than on weekdays, and the number of senior staff is reduced at the weekend.Reference Aylin, Alexandrescu, Jen, Mayer and Bottle 10 , Reference Mohammed, Sidhu, Rudge and Stevens 11 The number of specialist staff is also reduced, especially operation-specific theatre nurses, anaesthetists and recovery nurses.
In our study, all weekend elective tonsillectomies were carried out by senior medical staff, either consultants or associated specialists. This may have been a contributing factor to the results of the study. However, the UK National Prospective Tonsillectomy Audit suggested that grade of doctor is not associated with post-tonsillectomy haemorrhage rate,Reference Brown, Ryan, Yung, Browne, Copley and Cromwell 1 which is corroborated by our results. Tonsillectomy may be affected by differing staffing levels or experience to a lesser degree than high-risk procedures, and this may account for our overall results. We suggest that the ‘weekend effect’ is therefore most apparent for high-risk procedures that require specialist expertise during critical phases.Reference Roberts, Thorne, Akbari, Samuel and Williams 14
The authors were unable to identify any further studies that had investigated the relationship between the day of tonsillectomy and procedure complications.
Study limitations
The data were collected retrospectively using electronic hospital records; the results are therefore limited by the quality of routinely collected hospital data. A relatively small number of procedures were performed on the weekend (6.00 per cent) compared to the weekdays (94.00 per cent); this represents a vast difference between the two groups, but is largely unavoidable.
A difference in case mix between patients operated on at the weekend and on weekdays was evident. We acknowledge that full statistical correction for operative technique, age, gender and operator experience was not carried out, and these may represent confounding variables. We are also unable to adjust for any selection biases with regard to which patients were selected for weekend surgery. Identifying low-risk patients for weekend surgery is sensible and widely undertaken, but this would reduce complications in itself. This may infer that meticulous pre-operative stratification is essential to planning schedules and has more of an effect on outcomes than day of surgery.
In addition to these limitations, we cannot be certain that we have captured all instances of secondary post-tonsillectomy haemorrhage. Any case of post-tonsillectomy haemorrhage where the initial procedure was carried out at the study hospital would not have been included if the patient re-presented to another emergency department. Similarly, if any cases were not appropriately coded, they would have not been captured by this study.
A further multicentre study with a longer study period and prospective methodology would minimise these limitations.
-
• Several studies have identified increased morbidity and mortality rates for elective general surgical procedures performed at the weekend compared with the weekday
-
• These increased rates have been found for both high- and low-risk procedures
-
• This study suggests that the risk of post-tonsillectomy haemorrhage is not increased by weekend surgery
-
• The findings provide evidence against the ‘weekend effect’ for low-risk procedures
The authors suggest that despite the limitations of this study, the p-value is of significant value to cautiously provide validity to the overall conclusion.
Conclusion
Our analysis suggests that there is no increased risk of post-tonsillectomy haemorrhage if the procedure is carried out on the weekend compared with a weekday.
