Hypernatremia Risk Factors And Prognostic Impact In Burn Patients: A Case Control Study

Summary

Hypernatremia is associated with poor outcomes in critically ill patients. Hypernatremia risk factors in burned patients are not well studied. We aimed to identify hypernatremia risk factors and to evaluate outcomes in burned patients admitted to our burns intensive care unit. A case control study was conducted in adult burned patients hospitalized between January 1st 2017 and December 31st 2019. Cases who developed hypernatremia (>145 meq/L) during hospitalization were matched 1:1 with controls based on age and total burn surface area. There were 57 cases and 57 controls with a mean age of 41 ± 18 years. The majority of patients had major burns (n=99, 86.8%). The time onset of hypernatremia was seven days post burn. Compared to controls, the case group mostly consisted of transferred patients with longer time from injury to intensive care unit admission. Inhalation injury, mechanical ventilation, intravenous fosfomycin and colistin were associated with hypernatremia. Admission to the intensive care unit after six hours post-burn was the independent risk factor (OR=4.5). Hypernatremia was associated with longer length of stay and with higher mortality. We conclude that delayed management, inhalation injury, mechanical ventilation, fosfomycin and colistin administration are the main hypernatremia risk factors in burned patients

Introduction

Hypernatremia is a common electrolyte disturbance in hospitalized patients, especially in those who are critically ill.¹ Its incidence is around 2% and it is common in patients with neurological diseases and endocrine or renal diseases.² In critically ill patients, this incidence increases to 4.3% of patients with a higher mortality.³ In burn patients, hypernatremia incidence is even higher, ranging from 9.9%⁴ to 24.4%.⁵

The pathophysiological mechanisms of hypernatremia are related to fluid and sodium imbalance.⁶ In critically ill patients, hypernatremia can be multifactorial but its development is either a result of sodium overload or a loss of free water.⁶

In spite of water loss related to local lesions, burns induce systemic damage that seriously alters homeostasis. Systemic changes induce a hypercatabolic state with a greater risk of water balance disorders, 6 which might worsen patient outcomes. In burned patients, hypernatremia incidence and prognostic impact is reported in several studies but its risk factors are not well reported.

This study aimed to identify hypernatremia risk factors and to evaluate outcomes in burned patients admitted to our burns intensive care unit.

Patients and methods

This study was conducted at a 20-bed burns intensive care unit (burns-ICU) at a university teaching hospital in Tunis. Eligible patients were adult (age ≥18 years) burned patients who were admitted between January 1st 2017 and December 31st 2019. We have considered different sources that cause burns (fire/flame, electrical, chemical agents, diffuse epidermal exfoliation leading to toxic epidermal necrolysis).

Patients with a referral letter with recorded resuscitation fluid volumes prior to admission were eligible. Patients with concomitant trauma were excluded from the study.

Cases were patients who presented hypernatremia during hospitalization, defined as a serum sodium concentration greater than 145 meq/L. The cases were matched 1:1 with controls based on age and the initial assessment of total burn surface area (TBSA) calculated by the Wallace rule of nines.

If burns covered more than 20% of total body surface area, the patient was considered to have major burns. The following data were recorded: age, sex, preexisting comorbidities, burn mechanism and admission weight. Burn severity upon arrival was evaluated according to the abbreviated burn severity index (ABSI), Baux score (BS) and unit burn standard (UBS).

Inhalation injury was evoked in the presence of carbonaceous sputum, singed hairs, facial burns, and in the case of smoke exposure in a closed space. Patient management before hypernatremia diagnosis was assessed for the following: perfusion received to initial fluid resuscitation, mechanical ventilation, antibiotics, vasopressors and diuretic administration.

In patients transferred via the hospital emergency room, Parkland formula based on the burn size and body weight of the patient was applied as a first approximation of required fluid administration rates. Thereafter, fluid resuscitation was adapted in order to meet a target urine output of 0.5 to 1 ml/Kg/h and a target mean arterial pressure >65 mmHg.⁷

Statistical analysis was performed using SPSS software (SPSS Inc, Chicago, Illinois, United States). Univariate descriptive data were reported as counts and percentages for categorical variables, and the mean (standard deviation ± SD) or median [interquartile range] (IQR) for continuous variables.

To explore hypernatremia risk factors, univariate analyses were performed using Pearson Chi square or Fisher’s exact test for categorical data and the independent T-test or Mann–Whitney U tests for continuous variables. Cutoff values for predicting hypernatremia were calculated using the best values of sensitivity and specificity for each factor based on the area under the receiver operating curve (ROC). Binary logistic regression analysis was then performed using all variables that had p < 0.05 on univariate analyses. Odds ratios (OR) and 95% confidence intervals (CI) were calculated. Statistical significance was accepted at p <0.05.

The local institutional ethics committee approved this study.

Results

The study group consisted of 57 cases and 57 controls with a mean age of 41 ± 18 years. There were 69 males and 45 females (sex ratio=1.5). The majority of patients had no medical history (83.3%, n=95). Patients were hospitalized with thermal injuries in 105 cases (92.1%), electrical injuries in 6 cases (5.3%) and toxic epidermal necrolysis in three cases (2.6%). Seventy-six patients (66.7%) were transferred from another institution. Mean TBSA was 43 ±20% and the majority of patients had major burns (86.8%, n=99). Twenty-three patients (20.2%) had inhalational injury.

Among the cases, hypernatremia was diagnosed within seven days [1,9] post burn. Five patients, all transferred patients, were hypernatremic upon burns-ICU admission. In the cases, natremia was 147 mmo/L [147,152]. The median duration of hypernatremia was 4 days [4,5].

The control group was comparable to cases in terms of mean age, gender, comorbidities, burn mechanism and burn severity scores. Compared to the controls, the cases were mostly transferred patients who had a longer time from injury to burns-ICU admission. Inhalation injury was significantly more frequent in patients with hypernatremia (p=0.03).

Length of burns-ICU stay was longer in the cases (15 days [8,19] compared to 9 days [6,16] in the control group (p= 0.014)). In the study group, mortality was 83.3%. The intra hospital mortality rate after hypernatremia was 94.7%, versus 71.9% in the control group (p=0.001). Amongst the cases, natremia was not different between survivors and non-survivors (148 [147.152] versus 147 [147,152] meq/L). Univariate analysis and comparison of clinical characteristics of cases and controls are described in Table I.

Table I. Univariate analysis and comparison of clinical characteristics between cases and controls.

We found that admission to burns-ICU after six hours post-burn had the best sensitivity (72%) and specificity (62%) to predict hypernatremia. The area under the ROC curve was 0.74 (Fig. 1).

Fig. 1. ROC curve of time of intensive care admission for the prediction of hypernatremia.

Regarding the different treatments provided to patients, there was no significant difference between groups in fluid therapy solution or the first day total fluid requirement. Hypernatremia was more frequent in ventilated patients (p=0.03), in patients treated with intravenous fosfomycin (p=0.004) or colistin (p=0.004). The control group was comparable to cases in terms of the need for catecholamines (Table II).

Table II. Univariate analysis and comparison of management between cases and controls.

On multivariate analysis, only admission to the burns-ICU after six hours post-burn was independently associated with hypernatremia. The OR increased to 4.5 (95% CI [1.5, 13]) with maintained statistical significance (Table III).

Table III. Multivariate analysis of factors predicting hypernatremia.

Discussion

The main findings of this study suggest the importance of initial management of burns. Hypernatremia risk factors were transfer from another institution and an admission within six hours post burn. Hypernatremia may reflect hypovolemia occurring with sodium loss and relatively greater loss of body water.⁶

In addition to local lesions that cause increased evaporative water, burn injuries may have systemic effects. Systemic inflammation and subsequent vasodilatation and capillary leakage occurred on large burned surfaces.⁸ These changes coupled with fluid loss from the burn wound compromise water balance, and hypernatremia may be one of the consequences of body water loss. Consequently, fluid therapy is a mainstay in the resuscitation of burn and must be initiated as quickly as possible.

In the present study, we found that admission after six hours post-burn was an independent risk factor of hospital-acquired hypernatremia. Furthermore, hypernatremia occurred especially in patients transferred from another institution, which gives rise to the presumption that the initial fluid resuscitation had not been optimal. We also suggest that initial fluid resuscitation may have an impact on hypernatremia occurrence since we noted that the first seven days post-burn were the time onset of hypernatremia, as shown in previous studies.^4,5 Patients should be managed correctly while waiting for transfer to the burns-ICU, which indicates that non-burn specialist clinicians have to improve and update their knowledge and skills about burns management.

Lactated Ringer’s was the prescribed solution in almost all patients, conforming to the current guidelines.⁹

Therefore there was no significant difference between groups in fluid therapy solution. Fluid resuscitation must be directed with hemodynamic endpoints such as a target urine output and a target mean arterial pressure. Otherwise, hypernatremic state may involve a gain of sodium secondary to excessive fluid intake or prolonged fluid removal strategy.^10,11

As we conducted a case control study with a matching on age and TBSA, the study groups were comparable in these parameters. Severely burned patients, with subsequent fluid loss, were at risk of developing hypernatremia. ^4,5 Stewart et al.⁴ showed that patients with severe hypernatremia (sodium>150 mmol/L) had significantly higher burn sizes (45% versus 8%).

Approving the same definition of hypernatremia (sodium>145 mmol/L) as our study, Lam and Minh5 found that development of hypernatremia was associated with largely burned patients (TBSA ≥40%) with an OR=8.46. In our study, we collected mainly majorburned patients with a mean TBSA of 43%. This matching factor may be a confounding factor in previous studies, in which the difference in hypernatremia incidence may be related to burn severity.

In contrast with these findings, Namdar et al.’s¹² study, including patients with a mean TBSA of 26±12%, did not show a difference in ABSI or TBSA between patients without or with hypernatremia (>146 mmol/l).

Several studies showed a significant association between hypernatremia and mechanical ventilation and inhalation injury.⁵ Our results confirmed the contribution of these factors to the hypernatremic state. In critically ill patients, the risk of hypernatremia may be related to compromised capacity to maintain water intake due to sedation, mechanical ventilation and altered mental status.³ This risk may impel intensivists to be more attentive to the prevention of hospital-acquired hypernatremia and maintaining appropriate fluid and sodium balance.

Ventilator days, which can be associated with severe burns and inhalation injury, were associated with hypernatremic state in critically ill patients.^1,3

In critically ill patients, hypernatremia can have an iatrogenic origin. Fosfomycin is one of the sodium-rich antibiotics and each gram of IV fosfomycin contains 0.33 grams of sodium.¹ So hypernatremia is one of the side effects of fosfomycin, especially when given in high doses or for prolonged periods.¹³

Although sodium overload related to fosfomycin administration is clear, colistin is not known to provoke hypernatremia. We suggested that the contribution of colistin might be related to sepsis independently to this antibiotic. Indeed, sepsis patients are more likely to develop hypernatremia when compared with other ICU patients.^14,15 Hypernatremia complicates 4.7% of septic patients.¹⁵ In septic patients, hypernatremia may be related to renal water loss secondary to acute renal failure and to insensible losses caused by fever.^14,16

As in the present work, several studies have established the association between hypernatremia and poor outcomes.^1,5,17

Although mortality in hypernatremic patients (94.7%) is higher than that reported in other studies (30% to 48%),^1,5 the interpretation of this difference should consider the other mortality risk factors. On the other hand, the imputability of hypernatremia severity and sodium variability may be associated with an increased mortality in severely burned patients and may reflect acquired sodium disturbances.¹⁶

The strengths of our study exist in the similarity of cases and controls after matching, which limited the confounding factors. There were several limitations in the study. First, this was a monocentric retrospective study and all data were collected from medical records. Second, it included a relatively small number of patients. Also, we did not collect other hypernatremia risk factors such as sepsis.

Conclusion

Hypernatremic state was associated with higher mortality and length of stay. We suggest that initial management by non-burn specialist clinicians with subsequent delayed admission to the burn intensive care unit was the main risk factor for hypernatremia. These findings impel specialists to develop and provide medical training for non-specialists in order to improve burn outcomes.