Determinants and time to blood transfusion among thermal burn patients admitted to Mulago Hospital

Due to the demonstrable complications of transfusion among burns patients in concert with the improved patient outcome on restrictive blood use trials, the role of blood transfusion in burns has undergone re-evaluation over years [1, 2]. These data, have led to recommendations that blood transfusions be used only when there is an apparent physiologic need [3] based on the patients vital signs, estimation of blood loss and evaluation of blood volume, clinical and laboratory evaluation of end organ perfusion.

Current transfusion protocols use a specific (trigger) level of haemoglobin (Hb) or haematocrit (HCT) that dictates when to transfuse packed red blood cells (PRBCs). An acceptable strategy for transfusion of burn patients has not been specifically identified [4]. There is no one ‘common trigger’ but with on-going research, different triggers are adopted. This has changed from the traditional Hb <10 g/dL in the 1960s to <7 g/dL in the 1990s to a range of haemoglobin level of 6–8 g/dL [5–7], which is not used independently but alongside other physiological parameters to decide the need to or not to transfuse the patient.

PRBs transfusion policies at several U.S. burn units surveyed, found that haemoglobin as low as 6 g/dL and haematocrit as low as 15% were acceptable for healthy patients needing limited surgery. The highest haemoglobin considered as a transfusion trigger for critically ill patients with extensive burns and/or burns with cardiopulmonary compromise was 10 g/dL. For the stable burn patients’ recommendations to transfuse were at haemoglobin of ≤8 g/dL for non-critically ill and without cardiopulmonary compromise and for critically ill patients and/or those with cardiopulmonary compromise, transfuse at haemoglobin of ≤10 g/dL [5].

Kwan et al’s study implanting a restrictive transfusion policy in burned children showed that a haemoglobin threshold of 7 g/dL had no more adverse outcomes versus a traditional transfusion trigger of 10 g/dL. In addition, costs incurred to the institution were significantly less [4]. Similar results from other studies are shown [6, 7].

Other strategies to minimize transfusion requirement included minimizing volume of blood lost during surgeries by early wound excision within the first 7 days, the use of the modified tumescent surgical technique using subcutaneous adrenaline and adrenaline saline soaked non adherent pads to reduce intra-operative and total blood transfusion requirement as well as continuous tourniquet application during tangential excision on extremities in burn patients [8–11].

Most studies have defined the transfusion trends in burn patients, focusing on the relationship between % TBSA burn and transfusion needs [12–14].

The rate of transfusion increases with increasing % TBSA burn but with variations for the same burn extents in different study populations e.g. Graves et al. patients with >10% TBSA burn received an average of 19.7 units of blood [13],while for Vasko et al., their patients with >10% TBSA burn required an average of 8.94 units per patient [14]. However this conclusion overly generalizes patients as >10% could mean 11, 25, 70%. For patients with >30% TBSA, the mean transfusion requirement was 17 units [14]. Palmieri et al. found that, on average, 13.7 ± 1.1 units of PRBCs were transfused per patient with >20% TBSA, for burns of >50% TBSA, 30 units of blood were transfused per patient [12]. Approximately 5.7% of patients with <10% TBSA burn, 21% with 11–20% TBSA burn, 39% with 21–30% TBSA burn, and 62% of patients with >30% TBSA burn required a transfusion [15]. This may be attributed to increased surgical blood loss from extensive excision and grafting and also from worsened illness severity leading to impaired erythropoiesis.

Most frequently, on-going blood loss (22%), anaemia (20%), hypoxia (13%) and cardiac disease (12%) are the reasons for transfusion in thermal burn patients, with inhalation burns influencing the decision to transfuse blood in 34% [16]. Acute respiratory distress syndrome (ARDS), and age in a survey of trends of blood transfusion among surgeons were added as other factors increasing transfusion rates apart from the above mentioned [16]. Other reasons for blood transfusion in thermal burn patients were, critical illness extent, other trauma, age, sepsis, need for grafting and, TBSA [16].

Mann et al. found that there was no association of blood transfusion requirement with patient age, timing of the 1st surgical procedure from the time of injury or the length of patient stay in hospital [6].

Considering the available literature on blood transfusion in thermal burns patients, it is clear that little is known about other patient characteristics predisposing to transfusion in burn patients except/other than low haemoglobin levels and % TBSA of burns. However one does not need to labor to explain that factors like, haemodynamic status, nutrition status, may also influence the decision to transfuse at a higher Hb level. Amidst all this, the National Institute of Health consensus conference concluded that no single measurement can replace good clinical judgment concerning the need for RBCs transfusion and this goes for both indication and timing of transfusion.

Generally, compared to other causes of trauma, patients with burn injury are transfused way later from time of injury. In trauma other than burns, the haemoglobin level obtained shortly after injury may not detect occult bleeding or may be confounded by crystalloid related haemodilution but a level <10 g/dL in the first 30 min of patient injury will correctly identify significant blood loss in 90% of trauma patients hence the early transfusion.

In burns, the drop in hemoglobin level is classified as either acute blood loss, well established as the major cause of anaemia in burn injury occurring within the first 1–2 weeks of burn injury, or anaemia of critical illness which occurs between operative events, during wound healing and throughout the resolution of the acute phase of injury and accounts for up to 70% of cases that are transfused. Each of these two categories differs in aetiology from the other [17].

On average, burns patients are transfused 5.2 ± 0.3 days after admission [12]. This has been interpreted that direct thermal injury and sequestration of red blood cells doesn’t have a profound effect on the overall Hb level, otherwise transfusion would be done much earlier as is the case in other causes of trauma.

In another study, burn patients on average were transfused within 28.8 days from time of injury which is normally after all surgeries and relevant procedures required are done [16] however this contradicts the practice of early wound excision, one of the common surgeries performed in burn patients evidenced to decrease blood loss if performed in the first 24 h of injury [9, 18, 19].

Among critically ill patients including burns patients 50% of transfusions occurred in the first 5 days of ICU stay [20]. In the Anemia and blood transfusion in critically ill patients (ABC) study 70% of all transfused patients were transfused in the first 2 ICU days [21] which is still within the first week of admission.

In Mulago Hospital, many burns patients get transfused during their course of treatment however; there is reported scarcity of blood as a resource in across the country mainly due to low responses by citizens to blood donation. This could cause a delay or failure to transfuse patients who may need it.

The transfusion rate and indications of blood transfusion are not documented. The threshold of transfusion possibly a haemoglobin level of <8 g/dL is a typical consideration in our setting. It also varies from one patient to another when clinical and physiological assessments of these patients are considered by the specialist or attending clinician for that matter, as is the case in other burn centers in different countries. Literature on these predictors for the need to transfuse are not well established to allow one to foresee the need for transfusion right from the point of admission in Mulago hence the need for this study to address this gap in information.

We carried out a prospective cohort study for a period of 5 months only in Mulago National Referral. Mulago National Referral hospital being one of two such hospitals in the country with a representative population from across the country some self-referred others transferred in from peripheral and far off hospitals across the country. We had no conflict of interest and research funding was by the corresponding Author. We recruited 112 thermal burn patients admitted to the Surgical Unit of the Accidents and Emergency Department within 72 h from the time of the burn injury. Patients with a history of blood transfusions during the episode of injury but prior to admission were excluded.

All patients received standard burns care as per the Burns Unit of Mulago Hospital protocol which included, among others, resuscitation according to the Acute Trauma and life Support (ATLS) protocol, wound care, Fluid replacement according to the parkland formula, pain management, electrolyte replacement and nutritional rehabilitation, infection control, and counseling. The wound care however, did not involve early wound excision. Each patient was prospectively followed up while on the ward for a period of 28 days or until discharge, death or loss to follow up whichever came first. A pretested interviewer administered questionnaire was used to collect data on demographics, history of injury, co-morbidities (respiratory, cardiac disease, hematological disorders and epilepsy among others) and clinical examination findings on severity of burns and systemic examination from consenting participants.

Results from serial laboratory analyses carried out once a week for CBC, LFTs levels at admission and during ward care to diagnose for anaemia, hypoalbuminaemia, and sepsis (determined clinically and white blood cell count of <4000 or >12,000 cell/uL on laboratory analysis) were documented. During the process of following up study participants, we took note of those patients that got blood transfusion; the time of blood transfusion from time of injury, and their Pre transfusion Hb was recorded. There is no blood transfusion protocol in Mulago and the decision to transfuse was collectively made by a team of attending plastic surgeons, residents and general practitioners during ward rounds.

We also reviewed the patients’ files for any other indications of blood transfusion according to the attending team and also if these patients had had any surgery during the study period. We compared the variables above between the patients that received blood with those among patients that did not receive a unit of blood during the period of study. The cut off hemoglobin levels of 11.9–10, 9.9–8, 6–7.9 g/dL and <5.9 were used to stratify patients on basis of HB levels classified into severity groups. The data was checked for errors and corrected, and it was entered in a software package, EPIDATA and exported to a STATA for analysis. Patient confidentiality was maintained and patients’ details were not included in the data set. The hard copies of the questionnaire were kept under lock and key while the soft copy of the data was password protected. Comparisons of demographics and other baseline characteristics of these thermal burn patients were made between patients who received blood transfusion and those who did not receive blood transfusion.

The primary outcome for this analysis was blood transfusion. The study used survival analysis to estimate the time to blood transfusion among patients with different cut offs of TBSA and Hb levels. Kaplan–Meier survival methods were used and time of start of risk was at the date of injury. Patients who were not transfused and alive were censored at day 28. Patients who died or those discharged or lost to follow-up before day 28 were followed up to the date of death, date of discharge or last date seen at the ward whichever came first. A log rank test was used to test for equality of survival for blood transfusion. In a multivariate Cox Proportional Hazard model, the study attempted to identify which factors were associated with blood transfusion among thermal burn patients. All analysis was carried out using STATA 12.1 and a p value of p < 0.05 was considered as of statistical significance.