Mortality in patients receiving prolonged invasive mechanical ventilation time in the emergency department: A retrospective cohort study.

Background: Patients waiting for intensive care unit (ICU) admission cause emergency department (ED) crowding and have an increased risk of mortality and length of stay (LOS) in hospital, which increase the hospitalization cost. This study aimed to investigate the correlation between mortality and invasive mechanical ventilation (IMV) time in patients in the ED.
Methods: A retrospective cohort study was conducted in patients who received IMV in the ED of Ramathibodi Hospital. The correlation between mortality at 28 days after intubation and IMV time in the ED was analyzed. The cutoff time was analyzed to determine prolonged and nonprolonged IMV times. ICU ventilation time, length of ICU stay, and LOS in the hospital were also analyzed to determine their correlations between IMV time in the ED.
Results: In this study, 302 patients were enrolled, 71 died, and 231 survived 28 days after receiving IMV in the ED. We found that the duration of >12 h of IMV in the ED increased the 28-day mortality rate by 1.98 times (P = 0.036). No correlations were found between IMV time in the ED and ventilation time in the ICU, length of ICU stay, and LOS in the hospital.
Conclusion: More than 12 h of IMV time in the ED correlated with mortality at 28 days after initiation of IMV. No associations were found between prolonged IMV time in the ED with ventilation time in the ICU, length of ICU stay, and LOS in the hospital. Copyright:

INTRODUCTION

Emergency department (ED) crowding is a common problem in public hospitals and a significant public health problem in the Americas, Europe, Asia, Australia, and Africa[1] that leads to various disadvantages, significantly diminished quality of patient care. Owing to ED crowding, patients must wait longer to be attended by a physician and stay longer in the ED, which increases their risk of mortality.[2] A previous study showed that patients who received emergency care in crowded ED had increased 28-day mortality of 1.34 and that ED crowding affected the management of diseases where time is an important variable.[3] ED becomes crowded because of the high number of patients who require emergency service and delayed patient admission.[4] The solution to this problem is to improve the patient care process in the ED and the hospital system as a whole.[56]

Patients receiving invasive mechanical ventilation (IMV) in the ED are considered critical cases. A previous study showed that the 28-day mortality after admission to the intensive care unit (ICU) was 30%.[7] Therefore, patients requiring IMV should receive care in the ICU or intermediate ICU, but delay ICU admission may occur due to the hospital capacity or ICU bed not being available. Usually, these patients should be admitted to the ICU within 6 h after emergency room administration.[8] Patients with delayed admission have an increased risk of mortality of 1.4 times. Moreover, the length of stay (LOS) in hospitals is increased by 1.5 times[9] because of limited ICU resources.[10] Patients waiting for ICU admission cause ED crowding and have an increased risk of mortality and LOS in hospitals, increasing hospitalization costs.[11]

In Ramathibodi Hospital, more than 55,000 patients visit ED service/year, of whom more than 8,000 patients were admitted as inpatients. More than 50 patients/month require IMV, and more than 500 patients/month had LOS for more than 24 h in ED.

This research study aimed to determine the correlation between IMV time in the ED and mortality rate, including LOS in hospitals, to improve the quality of patient care and fluidity of the patient care process in both the ED and ICU.

METHODS

Study design

This retrospective cohort study was conducted in the ED of Ramathibodi Hospital. Patients (>15 years old) with respiratory failure who received IMV in the ED and did not meet the exclusion criteria were evaluated. The research was conducted between October 1, 2019, and May 30, 2020.

Methods

Data were collected from RAMA electronic medical records in five parts as follows: part 1, baseline characteristics such as age, sex, work shift, and emergency severity index (ESI) triage level; Part 2, intubation data such as place of intubation, history of intubation, intubation protocol, and indication for intubation; Part 3, data of illness, such as vital signs at first presentation, underlying disease, and diagnosis; Part 4, prognostic severity scores such as acute physiology and chronic health evaluation II (APACHE II) score on the 1st day of receiving IMV in the ED (ED APACHE II score), first 24 h after admission in the ICU (ICU APACHE II score), difference between the ED and ICU APACHE II scores (delta-APACHE II score), sequential organ failure assessment (SOFA) score on the 1st day of receiving IMV in the ED, SOFA score in the first 24 h after ICU admission, and difference between the ED and ICU SOFA scores (delta-SOFA score); and Part 5, data on time and result of treatment, such as ER arrival time, intubation time, time from ED arrival to ventilation, inpatient department/ICU arrival time, total ventilation time in the ED, extubation time, total ICU ventilation time, ICU discharge time, total length of ICU stay, hospital discharge time, total LOS in hospital, discharge status, and 28-day mortality. Data were collected from patients who met the inclusion criteria and then analyzed, and the research results were summarized following the research objectives.

Participants

The inclusion criteria were patients with respiratory failure (>15 years old) who received IMV in the ED of the Ramathibodi Hospital. Trauma patients, patients intubated because of cardiac arrest, patients with a past and present history of tracheostomy, patients referred from and to other hospitals, postcardiac arrest patients, patients with incomplete medical records, and patients who refused medical care were excluded from the study.

Sample size calculation

According to Hung et al.,[9] IMV time in the ED of >4 h was considered prolonged. A pilot study was conducted on 40 patients and obtained the following values: P (outcome/exposure) of 0.17, P (outcome/exposure) of 0.19, alpha of 0.05, and a beta of 0.2. The study sample of 554 patients included 277 and 277 previously exposed and unexposed to IMV.

Statistical analyses

The patients’ baseline characteristics were evaluated using descriptive statistics and presented as frequency and percentage for categorical data and mean and standard deviation for continuous data. The Chi-square or Fisher exact test was used to compare the categorical data, and the Student t-test was used to compare continuous data. The analysis results using a multivariate logistic regression model are presented as odds ratio (95% confidence interval [CI]). Significant factors were defined as those with P < 0.05. All statistical analyses were performed by Stata 16 (StataCorp, College Station, TX, USA).

RESULTS

At baseline, of 302 patients, 155 were male (51.32%) and 147 were female (48.68%) [Figure 1]. The mean age in both groups was 72.23 years. We found no significant differences between the two groups when considering 28-day mortality after IMV in the ED.

Figure 1

Research method diagram

The patients were divided into five levels according to the ESI triage as follows: ESI level 1, 41 patients (13.58%); ESI Level 2, 222 patients (73.51%); and ESI Level 3, 34 patients (11.26%). Most patients received medical service during the evening shift. No significant differences in ESI triage level and work shift were found.

The indications for intubation in 251 patients (83.39%) were oxygenation and ventilation failure, and sedative agents were used as an intubation protocol in 176 patients (58.28%). When considering 28-day mortality after IMV in the ED, we found no significant differences in intubation and intubation protocol indications.

The incidence rates of underlying diseases, namely hypertension in 165 patients (54.64%), diabetes in 31.79% of the patients, chronic kidney disease in 26.49%, and cancer in 25.50%, were similar. Of the patients, 146 (48.34%) and 40 (13.25%) were diagnosed with respiratory tract infection and pulmonary edema, respectively.

The patients were evaluated using the APACHE II and SOFA scores on the 1st day of receiving IMV in the ED. The mean ED APACHE II score was 21.79, and the mean ED SOFA score was 5.62. When the patients were reevaluated after admission for 24 h in the ICU, the mean ICU APACHE II and SOFA scores were 20.08 and 5.71, respectively. Statistically significant differences were found between the ED and ICU APACHE II and SOFA scores.

The mean total length of ED stay and ventilation time was 21.88 and 13.02 h, respectively. The mean total ICU ventilation time and length of hospital stay were 9.36 and 20 days, respectively. No significant differences were found between the two groups. However, the mean length of hospital stay and total length of ICU stay were 20 days (9.92%) and 12.78 days (13.66%), respectively, with statistically significant differences between the groups [Table 1].

Table 1

Patients’ baseline characteristics

	Total	Died (n=71)	Survived (n=231)	P
Age (years)	72.23±4.16	71.13±14.88	72.56±13.96	0.455
Sex, male (%)	155 (51.32)	36 (50.70)	119 (51.52)	1.000
ESI triage (%)
Level 1	41 (13.58)	11 (15.49)	30 (12.99)	0.940
Level 2	222 (73.51)	51 (71.83)	171 (74.03)
Level 3	34 (11.26)	8 (11.27)	26 (11.26)
Shift work (%)
Morning (8:00 am-4:00 pm)	119 (39.40)	25 (35.21)	94 (40.69)	0.275
Evening (4:00 pm-00 am)	128 (42.38)	36 (50.70)	92 (39.83)
Night (00 am-8:00 am)	55 (18.21)	10 (14.08)	45 (19.48)
Intubation protocol (%)
RSI	76 (25.17)	18 (25.35)	58 (25.11)	0.466
Awake	50 (16.56)	15 (21.13)	35 (15.15)
Sedation	176 (58.28)	38 (53.52)	138 (59.74)
Indication for intubation (%)
Oxygenation/ventilation failure	251 (83.39)	58 (82.86)	193 (83.55)	0.913
Upper airway obstruction	2 (0.66)	0	2 (0.87)
Maintain/protect airway	48 (15.95)	12 (17.14)	36 (15.58)
Underlying disease (%)
Diabetes mellitus	96 (31.79)	21 (29.58)	75 (32.47)	0.771
Hypertension	165 (54.64)	34 (47.89)	131 (56.71)	0.220
Chronic kidney disease	80 (26.49)	16 (22.54)	64 (27.71)	0.444
Ischemic heart disease	48 (15.89)	9 (12.68)	39 (16.88)	0.462
COPD	32 (10.60)	10 (14.08)	22 (9.52)	0.276
Asthma	6 (1.99)	0	6 (2.60)	0.342
Chronic heart failure	14 (4.64)	5 (7.04)	9 (3.90)	0.330
Old CVA	60 (19.87)	9 (12.68)	51 (22.08)	0.091
Chronic lung disease	12 (3.97)	1 (1.41)	11 (4.76)	0.306
Active pulmonary TB	8 (2.65)	1 (1.41)	7 (3.03)	0.686
Old pulmonary TB	16 (5.30)	1 (1.41)	15 (6.49)	0.130
Cirrhosis	25 (8.28)	6 (8.45)	19 (8.23)	1.000
Autoimmune disease	8 (2.65)	3 (4.23)	5 (2.16)	0.397
Malignancy	77 (25.50)	25 (35.21)	52 (22.51)	0.042
Diagnosis (%)				0.586
Respiratory tract infection	146 (48.34)	31 (43.66)	115 (49.78)
Pulmonary edema	40 (13.25)	7 (9.86)	33 (14.29)
Sepsis	33 (10.93)	9 (12.68)	24 (10.39)
Stroke	21 (6.95)	6 (8.45)	15 (6.49)
COPD with AE	10 (3.31)	1 (1.41)	9 (3.90)
UGIB	6 (1.99)	2 (2.82)	4 (1.73)
Seizure	3 (0.99)	1 (1.41)	2 (0.87)
ACS	2 (066)	0	2 (0.87)
Acute asthmatic attack	1 (0.33)	0	1 (0.43)
ED APACHE II score (mean)	21.79±8.92	24.17±9.18	21.06±8.73	0.009
ICU APACHE II score (mean)	20.08±8.49	24.97±9.50	18.58±7.57	<0.001
ED SOFA score (mean)	5.62±2.80	6.86±2.99	5.24±2.64	<0.001
ICU SOFA score (mean)	5.71±2.91	7.44±3.09	5.19±2.64	<0.001
Delta-APACHE II score	−1.71±6.48	0.80±7.28	−2.48±6.02	<0.001
Delta-SOFA score	0.09±1.77	0.58±2.12	−0.06±1.62	0.008
Total ED ventilation time (h)	13.02±19.01	11.40±14.10	13.52±20.28	0.410
Total ED stay (h)	21.88±29.02	22.72±29.95	21.62±28.78	0.779
Total ICU ventilation time (days)	9.36±9.81	9.23±8.51	9.40±10.20	0.896
Total ICU stay (days)	12.78±12.05	9.92±8.57	13.66±12.81	0.021
Total hospital stay (days)	20.00±22.64	10.61±8.77	22.89±24.73	<0.001

ESI: Emergency severity index, COPD: Chronic obstructive pulmonary disease, TB: Tuberculosis, ICU: Intensive care unit, ED: Emergency department, APACHE-II: Acute Physiology and Chronic Health Evaluation II, SOFA: Sequential organ failure assessment, AE: Acute exacerbation, UGIB: Upper gastrointestinal bleeding, RSI: Rapid sequence induction, CVA: Cerebrovascular accident

According to our statistical data analysis using multivariate logistic regression and patient groups divided according to IMV time, with longer 2 h, the odds of mortality of the patients with IMV time in the ED of >12 h at 28 days after intubation was 1.98 times higher than in those with ventilation times <12 h, and this difference is statistically significant [Table 2].

Table 2

Correlation between 28-day mortality and invasive mechanical ventilation time in emergency department

Total ED ventilation time (h)	OR (95% CI)	P
>4	0.60 (0.33-1.10)	0.101
>6	0.79 (0.44-1.44)	0.440
>8	1.11 (0.60-2.03)	0.737
>10	1.61 (0.87-2.99)	0.128
>12	1.98 (1.05-3.77)	0.036
>14	1.96 (1.02-3.76)	0.043
>16	1.49 (0.74-2.98)	0.254
>18	1.42 (0.67-3.03)	0.362
>20	2.23 (1.00-4.95)	0.048
>22	1.94 (0.81-4.62)	0.136
>24	1.91 (0.74-4.96)	0.183

OR: Odds ratio, CI: Confidence interval, ED: Emergency department

As shown in Table 2, IMV time >12 h was defined as the cutoff prolonged IMV time. It was used to determine the correlations of IMV time in the ED, ventilation time in the ICU, length of ICU stay, and length of hospital stay [Table 3].

Table 3

Correlation of emergency department invasive mechanical ventilation time, intensive care unit ventilation time, length of intensive care unit stay, and length of hospital stay

ED ventilation time	<12 h	>12 h	P
ICU ventilation time (days, mean±SD)	9.21±10.34	9.82±9.79	0.632
ICU stay (days, mean±SD)	12.82±12.33	12.51±10.56	0.840
Hospital stay (days, mean±SD)	20.01±23.99	19.34±19.03	0.819

ICU: Intensive care unit, SD: Standard deviation, ED: Emergency department

DISCUSSION

Our results show that IMV time of >12 h at 28 days after intubation correlated with statistically significant odds of increasing ICU ventilation time, length of ICU stay, and length of hospital stay, but IMV time <12 h showed no such correlations.

Many recent studies have investigated the cutoff time to define prolonged IMV in the ED. Hsieh et al.[12] found that time of >1 h defines prolonged IMV in the ED that increases the odds of mortality by 2.18 times (95% CI, 1.07–4.45, P = 0.03). Hung et al.[9] found that IMV time of >4 h increased the odds of mortality by 1.41 times (95% CI, 1.05–1.89; P = 0.024). Their study also showed a trend of increasing odds, consistent with the results of the studies as mentioned earlier, but the cutoff time was 12 h, which increased the odds by 1.98 times (95% CI, 1.05–3.77; P = 0.036), which is different from their result. Previous studies were conducted in hospitals with more ICU beds, unlike in Ramathibodi Hospital, where patients wait longer for an ICU bed. Moreover, Donal et al.[13] found that the waiting time for an ICU bed of >6 h increased the length of hospital stay (mean, 6 days) as compared with <6 h (mean, 7 days; P < 0.001), which is related to the finding of Hung et al. that a waiting time for an ICU bed of >4 h increased the length of hospital stay to >30 days, with an odds of 1.56 times higher (95% CI, 1.07–2.27; P = 0.020). For this study, when considering a duration of >12 h for IMV in the ED, we found no correlation with ventilation time in the ICU, length of ICU stay, and length of hospital stay.

Intensive care and continuous close monitoring are essential in patients with critical conditions, especially those with acute respiratory failure. Delayed ICU admission shows evidence of the poor outcome.[11] ED crowding leads to ineffective patient care. Inadequate ICU capacity is one reason for patients with critical conditions’ more extended ED stay, especially those receiving IMV. Not delay mechanical ventilation in ED may improve the outcomes in critically ill patients.[14] This limitation leads to increased mortality, according to our study result.

The ED of Ramathibodi Hospital also faces this problem. From 2017 to 2017, approximately 500 patients had >24 h of ED stay, and >50 patients received IMV.

According to our study result, prolonged invasive ventilation time in the ED of >12 h increased the 28-day mortality rate significantly. However, our cutoff time (12 h) was longer than that in another study but represented the context of a more crowded hospital with lesser capacity. The findings from this study are useful in the allocation of ICU beds and the prioritization of patients for ICU admission.

Limitations

In this research, we collected patient data from a single-center Ramathibodi Hospital for only 1 year, which resulted in insufficient data for establishing strong statistical evidence.

CONCLUSION

More than 12 h of IMV time in the ED correlate with increased 28-day mortality, but prolonged IMV time in the ED showed no association with ventilation time in the ICU, length of ICU stay, and LOS in hospital.

Research quality and ethics statement

This study was approved by the Committee for Research, Faculty of Medicine, Ramathibodi Hospital Mahidol University (Approval # COA.MURA2019/945; Approval date September 28, 2019). The authors followed the applicable EQUATOR Network (http://www.equator-network.org/) guidelines, specifically the STROBE Guidelines, during the conduct of this research project.

Financial support and sponsorship

None.

Conflicts of interest

There are no conflicts of interest.