The reasons for applying MV in the ED vary, as do the application methods. MV may also affect the clinical course of the patient, while various clinical factors affect the duration of MV and its related outcomes. Therefore, it is challenging to accurately predict the duration of MV. A large-scale study on MV was performed by Esteban et al. in which the clinical features of 5,131 mechanically ventilated patients and MV methods were investigated. Their study also investigated the influencing factors of mortality among mechanically ventilated patients by dividing such factors into three categories: baseline conditions of patients at the time of MV, treatment-related factors, and events that occurred during MV [2]. Additionally, there have been other studies that investigated outcomes and mortality rates among patients according to their durations of MV and ICU stays; the duration of MV was often used as an outcome indicator for patients who were attached to ventilators [3].
The primary outcome in our present study was the total duration of MV, and we classified our patients into two groups based on a 7-day MV cutoff and compared their characteristics and clinical features. Previous studies have revealed slightly different average MV durations, although a study performed in 2002–2003 found that the average duration of MV is typically 3–4 days [13], and the novel term “prolonged acute MV” was subsequently created to refer to ≥ 96 hours of MV [14]. Because PMV is defined differently among studies, assigning a uniform cut-off is challenging. Our hospital recommends that emergency ICU stays not exceed 7 days, which we judged to be a meaningful cutoff for MV duration as well.
We found that the duration of MV varied according to our patients’ primary diagnosis or underlying disease. The mean duration of MV was ≥ 10 days in patients whose primary diagnosis was metabolic disorder, pneumonia, neurological disease, sepsis, or multiple trauma. Moreover, a high percentage of patients in the MV ≥ 7 days group had previously been diagnosed with stroke or dementia. It has been reported that a variety of factors influence the duration of MV when it is applied for neurological reasons; such factors include the mechanism of neurological abnormality, site of neurological injury, and state of consciousness [15, 16]. It was also reported that MV weaning in patients with neurological injuries requires an alternative means to maintain an open airway when subsequent positive pressure breathing is not required [17]; the numbers of patients in our study who underwent ≥ 7 days of MV due to neurological diagnosis and traumatic injury were 33 and 11, respectively, among whom 10 and 3, respectively, required tracheostomy. These proportions were similar to those of all patients requiring ≥ 7 days of MV who underwent tracheostomy regardless of their underlying conditions. However, unlike tracheostomy for positive pressure breathing owing to a respiratory disease, MV due to neurological reasons may be maintained solely for supporting the airway; thus, other methods such as early tracheostomy may also be required. Patients in our study who had stroke or dementia as an underlying disease required longer MV durations, which is believed to be the result of a higher risk of extubation failure owing to a decrease in cough, vomiting, and deglutition reflexes required for maintaining open airways in patients who are not fully conscious.
The MV < 7 days group included more patients with heart or kidney disease as the primary diagnosis or with chronic kidney disease as the underlying affliction. The patients may have been placed on MV due to respiratory distress caused by acute pulmonary edema arising from their condition, although acute respiratory distress was relieved within a short period after dialysis or diuretic administration.
MEWS is a known indicator of in-hospital outcomes of ED patients, including mortality and ICU admission [18]. In the present study, there was no difference in MEWS between the two MV groups, although the MV ≥ 7 days group comprised a higher proportion of patients who received norepinephrine. Because MEWS better reflects urgency rather than severity, patients in the MV ≥ 7 days group did not exhibit any initial vital sign abnormalities, but may have been more influenced by hemodynamic effects owing to the administration of sedatives.
Moreover, we found that patients in the MV ≥ 7 days group had higher CRP levels. A previous study found that patients with subacute conditions who had been on MV for approximately 2 weeks and who had low CRP levels had a higher rate of MV weaning than those with high CRP levels, although the difference was not significant [19].
The difference in blood albumin levels between the two groups was not statistically significant in our study; however, albumin was found to increase the risk of requiring MV for ≥ 7 days. Serum albumin level is an indicator of patient prognosis [20], and is associated with CRP levels, lactate levels, and fluid intake requirements. A lower albumin level is associated with an increased frequency of vasopressor use as well as a higher mortality rate. In our study, patients with albumin levels < 3.0 g/dL exhibited a higher risk of requiring MV for ≥ 7 days than those with albumin level < 3.5 g/dL. Xiao et al. reported that hypoalbuminemia increased the duration of MV, as albumin is associated with nutritional state, degree of inflammation, and pulmonary edema [21].
Creatinine levels were actually lower in the MV ≥ 7 days group, which was likely owing to pulmonary edema resolving soon after dialysis in patients in whom kidney disease as the underlying disease or primary diagnosis; this facilitated MV weaning. However, Pan et al. reported that acute kidney injury may be a significant risk factor for PMV when patients who are undergoing regular hemodialysis are excluded [22].
There was no difference in PaO2/FiO2 ratio between the two patient groups;
this value was found not to be a risk factor of PMV. Oh et al. reported that the ventilator days in pneumonia was related to PaO2/FiO2 ratio [23]. It has been reported that the PaO2/FiO2 ratio may differ between patients with acute respiratory distress syndrome who survive versus those who do not, although it does not necessarily serve as an indicator of mortality because of variations in MV settings, the patients’ condition, and patient posture. Our patients were placed under MV for various diagnoses and causes; thus, the PaO2/FiO2 ratio was not the sole factor considered for MV weaning. However, this ratio may be helpful in predicting the patient’s long-term treatment response, rather than serving as an indicator for the early prediction of MV weaning [24].
We found that the rates of reintubation, tracheostomy, ICU mortality, and 30-day mortality were higher among patients in the MV ≥ 7 days group, which is consistent with a previous study that showed that PMV and reintubation are associated with each other [25]. Other studies also found that longer MV durations increase the risk of mortality. Our PMV cutoff of 7 days was shorter than the 21 days defined by the NAMDRC Consensus Conference. However, our findings did not show significant differences in the PMV risk factors that were identified in previous studies.
The patient populations in existing studies are relatively homogenous, and usually comprise those already admitted to the ICU [11]. Our study aimed to predict the duration of MV during the early stage of ED treatment in patients who were placed on MV in the ED. Because we investigated the duration of MV among a diverse range of patients who were admitted to the ED, the study population included patients with a variety of internal diseases, trauma, and central nervous system disorders and were not limited to a certain affliction. As such, we were able to determine the effects of both the acute conditions and underlying diseases governing the duration of MV. Predictions of the duration of MV using hematologic test results and the initial clinical features of ED patients may not be as accurate as those based on patient progress after ICU admission. However, estimating the duration during early treatment stage in the ED or ICU could be helpful in setting early treatment goals and establishing long-term treatment plans based on such goals.
Our study had two major limitations. Because the same protocol for MV weaning was not applied to all patients, those with similar medical conditions may have experienced different durations of MV. This may be attributed to our study being preliminary, and a common MV weaning protocol will be applied to all patients going forward. Second, the patients did not all undergo the same laboratory tests; therefore, certain patient groups may have had no data for specific parameters, thereby resulting in a marked number of missing data points when assessing risk factors for PMV. Future investigations that apply the same protocols and basic tests to all patients ought to determine the risk of PMV more accurately.