Introduction

Tonsillectomy with or without adenoidectomy is one of the most common surgical procedures performed in children. The most common indications for adenotonsillectomy are airway obstruction resulting in sleep-disordered breathing or obstructive sleep apnoea (OSA), and recurrent infection.^{Reference Ingram and Friedman1} Even if surgery-related bleeding in paediatric patients is considered an uncommon event, adenotonsillectomy surgery has a non-negligible incidence of bleeding complications. The latter occur in about 2–7 per cent of cases, both early and delayed, and range from mild to life-threatening. The generally recognised causes of bleeding are surgical trauma, ineffective local haemostasis and local infections. Several studies have been conducted to determine the best strategy to screen paediatric patients at potential risk of bleeding during or after surgery.^{Reference De Luca Canto, Pachêco-Pereira, Aydinoz, Bhattacharjee, Tan and Kheirandish-Gozal2}

Post-operative haemorrhage following tonsillectomy can be classified as either primary/early (occurring within 24 hours of surgery) or secondary/delayed (occurring more than 24 hours after surgery). The incidence is variable, depending in part on how haemorrhage is defined and measured. Primary haemorrhage typically ranges from 0.2 to 2.2 per cent, and secondary haemorrhage between 0.1 and 3 per cent.^{Reference De Luca Canto, Pachêco-Pereira, Aydinoz, Bhattacharjee, Tan and Kheirandish-Gozal2–Reference Lowe, van der Meulen, Cromwell, Lewsey, Copley and Browne4} Secondary haemorrhage most commonly occurs 5–10 days post-operatively,^{Reference Liu, Anderson, Willging, Myer, Shott and Bratcher5} and is caused by premature separation of the eschar, which may be precipitated by an underlying infection or dehydration.^{Reference Johnson, Elluru and Myer6}

Post-operative haemorrhage usually stops spontaneously, but it sometimes requires a return to the operating theatre for haemorrhage control and, in rare cases, a blood transfusion. A minor bleeding episode increases the risk that a major bleeding episode will occur; it is therefore recommended that children who have experienced a bleeding episode be hospitalised overnight for observation.^{Reference Sarny, Ossimitz, Habermann and Stammberger7–Reference Lescanne, Chiron, Constant, Couloigner, Fauroux and Hassani9} These episodes can be life-threatening or, in rare cases, life-ending.^{Reference Windfuhr, Schloendorff, Sesterhenn, Prescher and Kremer10} A post-operative haemorrhage can cause difficulty in securing the airway by intubation, leading to an anoxic injury.^{Reference Subramanyam, Varughese, Willging and Sadhasivam3}

Children with bleeding disorders are at increased risk of post-operative bleeding complications, including delayed bleeding occurring up to two weeks after surgery. They therefore require close monitoring for a prolonged duration. Older children are at higher risk of post-operative haemorrhage compared with young children.^{Reference Liu, Anderson, Willging, Myer, Shott and Bratcher5,Reference Sarny, Ossimitz, Habermann and Stammberger7,Reference Hessén Söderman, Ericsson, Hemlin, Hultcrantz, Månsson and Roos11} Post-operative haemorrhage is also more common in patients who undergo tonsillectomy for chronic tonsillitis than for OSA.^{Reference Perkins, Liang, Gao, Shultz and Friedman12} This is because children with chronic tonsillitis have more scarring between the tonsil and the surrounding tissue, making removal more difficult.

Children with a haemorrhage of any cause can be categorised by severity into four classes, based on per cent loss of blood volume. Clinical features are typically used to estimate blood volume deficit.^{Reference Hazinski, Barkin and Barkin13,Reference Waltzman, Mooney, Major trauma, Fleisher, Ludwig and Henretig14} Many children presenting with class II haemorrhage will be in shock, and all children with classes III and IV haemorrhage will be in shock. The approach to haemorrhage is different depending on the haemorrhage class (Table 1).

Table 1. Haemorrhage management according to severity classification^{Reference Hazinski, Barkin and Barkin13,Reference Waltzman, Mooney, Major trauma, Fleisher, Ludwig and Henretig14}

Children compensate for blood loss by vasoconstriction and tachycardia, and may lose up to 30 per cent of their blood volume before hypotension is first observed.^{Reference Hazinski, Barkin and Barkin13,Reference Waltzman, Mooney, Major trauma, Fleisher, Ludwig and Henretig14} Tachycardia compensation is seen in a class II haemorrhage (15–30 per cent blood loss), while hypotension only appears when the haemorrhage is classified as class III (30–40 per cent blood loss). Blood loss also affects haematocrit and haemoglobin levels in the blood. Depending on the degree of blood loss and the amount of crystalloid replacement, the patient may present with significant anaemia. It can be difficult or impossible to quantify the blood loss from haemorrhage after tonsillectomy. Much of the blood can be swallowed, and vomited blood may be very difficult to estimate. Dizziness and orthostatic vital signs may suggest hypovolaemia in children who have been operated on.¹⁵

Currently, an approach based on a child's personal bleeding history is considered more cost-effective than a complete haemostatic screening, but there is no definitive way to predict which children will bleed. When a child loses blood, the physiological response is tachycardia, before blood pressure changes. Therefore, we wanted to investigate whether a child's heart rate is a reliable way to identify an initial change in blood volume.

We hypothesised that the relationship between heart rate and haematocrit could be associated with the amount of bleeding as opposed to the presence of bleeding, as post-tonsillectomy bleeding is readily apparent in the majority of cases. In addition, we expected that the decreases in haemoglobin and haematocrit could be predicted based on the changes in heart rate. We expected to find a relation between haemodynamic indices (heart rate) and laboratory indices (haemoglobin and haematocrit) in the setting of blood loss, in children. If such a relationship is found, it will provide a novel way to identify early bleeding events after surgical procedures by the simple measurement of heart rate.

This information might enable physicians to monitor patients more carefully after surgery. Parents or caregivers will be able to monitor their child post-surgery, at home, using basic medical tools, and if the child starts to bleed, the parents or caregivers will be able to provide them with appropriate medical care, long before their bleeding becomes severe. The information we expect to gather from this study might also provide us with a reliable means of monitoring children who undergo different kinds of surgical procedures or experience different traumas, and be used in different settings, whether in hospital or pre-hospital settings.