Readmission rate among survived patients with acute respiratory failure: 1-year study

Abstract

Context

Critically ill patients, especially those with respiratory failure associated with multiple comorbidities, are at risk of recurrent ICU admission and consuming a significant portion of medical resources.

Aim

To study the risk factors and common etiologies of readmission among surviving patients with acute respiratory failure in Mansoura University’s respiratory intensive care unit during the year 2023.

Settings and design

This was a retrospective cohort study.

Patients and methods

This study included all patients with acute respiratory failure who were admitted to Mansoura University’s respiratory intensive care unit (RICU), from January 2023 to December 2023. Deceased patients after initial admission or those with incomplete data were not included in the study. All data related to patient’s demographics, type of respiratory failure, type of respiratory support, associated comorbidities, length of ICU stay, and causes of readmission were collected from the registration system database.

Results

Eight-hundred-thirty (830) cases that were admitted to the Mansoura University’s RICU with respiratory failure in the year 2023 were enrolled in the study, and 84 cases (10.1%) of them were readmitted. COPD exacerbation was the most common cause of readmission (35.7%) followed by pneumonia (21.4%) and OSA exacerbation (20.2%). Results showed that the significant independent predictors of readmission were being male [AOR (95% CI): 1.8 (1.01–3.1), p: 0.046], having organ failure (renal/liver) [AOR (95% CI): 29.9 (7.9–113.4), p ≤ 0.001], and length of ICU stay more than 12 days [AOR (95% CI): 4.8 (2.6–8.6), p ≤ 0.001]. Also, the type of respiratory failure and the type of respiratory support received were significantly associated with readmission in the univariate analysis; however, they were insignificant in the multivariate analysis.

Conclusion

The rate of readmission was not high (only 10.3%). The most common causes of readmission were COPD exacerbation, followed by pneumonia and OSA exacerbation, respectively. Type of respiratory failure, male gender, the presence of other organ failure, and length of ICU stay are significant independent predictors of readmission in Mansoura University’s RICU during the year 2023. Therefore, higher-risk individuals should receive attention and careful assessment before discharge from the ICU to reduce the rate of readmission.

Trial registration

ClinicalTrials.gov identifier: NCT06291636.

Introduction

A significant portion of medical expenses in the intensive care unit (ICU) was consumed due to hospitalizations of patients with the most critical illnesses. Readmission to the ICU, either due to exacerbations of a previous disease in the initial admission or due to an unrelated new illness, results in further consumption of cost and medical resources and has been proposed as a marker for the quality of care [1].

The intensive care unit (ICU) readmission rates have been estimated at 4 to 14%, and different risk factors have been proposed by various studies [2]. In addition to the bad psychological impact for both patient and his family, readmitted patients mostly carry poor prognosis compared to newly admitted ones [3].

There is a developing interest in underlying the primary risk factors related to preventable ICU readmissions as well as in developing predictive tools and to help different ICU specialists to avoid premature ICU discharges [4].

It was found that criteria for ICU admission, discharge, and readmission vary from ICU to ICU and from country to country [5].

Although it is controversial, ICU readmission rate after initial discharge has been used as a marker of ICU and hospital performance [6, 7]. So, our aim was to measure the readmission rate, to study the risk factors, and to underline common etiologies of readmission among surviving patients with acute respiratory failure in Mansoura University’s RICU during the year 2023.

Patients and methods

Study setting

Mansoura University’s RICU, Chest Department.

Study design

Retrospective cohort study.

Patients

The study included all adult patients (≥ 18 years) with acute respiratory failure and who were admitted to Mansoura University’s RICU from January 2023 to December 2023.

Inclusion criteria

• Age 18 years or more

• Patient with acute respiratory failure (either type I or type II) who was treated with high-flow nasal cannula (HFNC), noninvasive ventilation (NIV), or invasive mechanical ventilation (IMV)

A readmission event was recorded if the same patient ID was present in the same year.

Exclusion criteria

• All patient with missed or incomplete data

Methods

From the database of Mansour University hospital registration, we collected the following data about all patients, with acute respiratory failure, who were admitted to Mansoura University’s RICU, Chest Department.

• Demographic data of the patients and diagnosis of initial admission

• Type of respiratory failure either hypoxemic respiratory failure (type I) or hypercapnic respiratory failure (type II)

• Type of respiratory support: High-flow oxygen therapy (e.g., HFNC), NIV, or IMV

• Associated comorbidities (e.g., hypertension, diabetes, congestive heart failure, chronic kidney disease, liver cirrhosis, malignancy)

• Critical illness severity score (SAPS II) [8]

The score is categorized into low severity if less than 20, moderate severity if the score is between 20 and 40, high severity if more than 40.

• Length of ICU stay and discharge status after initial admission

Sample size calculation

Elshafey and Hewidy (2014) found that 33.88% of patients who survived after ICU readmission were readmitted [1]. So, with a 95% confidence interval and margin of error 5%, the minimum calculated sample size would be 345. It was calculated using online calculator (Calculator.net) at https://www.calculator.net/sample-size-calculator.html?type=1&cl=95&ci=5&pp=33.88&ps=&x=Calculate.

Statistical analysis

Data was cleaned and tabulated via SPSS v. 26. Data was presented in the form of frequencies and percentages in the case of categorical data, while median and IQR were used for presenting continuous data after normality testing. Chi-square test and Fisher’s exact test were used for testing the significance of differences between the cases that were readmitted and the cases that were not readmitted. Univariate and multivariate binary logistic regression analyses were done using backward conditional method for determining the independent significant predictors of readmission and in hospital mortality. Variable entered in multivariate analysis of the predictors of readmission included type of sex, respiratory failure, and type of respiratory support, comorbidity, length of ICU stay, and discharge status, while age, comorbidity, type of respiratory support, length of ICU stay, and type of respiratory failure were entered in multivariate regression analysis of the predictors of in hospital mortality. The significance level was set at 0.05.

Results

The current study included 830 cases that were admitted to the Mansoura University’s RICU with respiratory failure. Eighty-four cases (10.1%) of them were readmitted to the RICU. Figure 1 is a flow chart showing the steps of inclusion in the study. Table 1 shows clinical characteristics of the patients and demonstrates the differences between the group of the cases that were readmitted and the group of cases that were discharged and not readmitted. Their median age of cases (IQR) was 61 years (18) with no significant differences between the groups (p: 0.220), with 56% of them were females with significant difference between the groups (p: 0.014). The most common cause of initial admission was pneumonia (50.6%), followed by COPD (34.2%), with significant difference between the groups (p ≤ 0.001). Most patients had no comorbidity (80.4%), with cancer, DM, and HTN being the most common comorbidities with nearly 5% for each disease, with significant difference between the groups (p ≤ 0.001).

Fig. 1

Flow chart of the study

Table 1 Clinical characteristics of the studied patients (N =  − 763)

Table 1 also shows that nearly about half of the patients had type 1 respiratory failure (47.2%) with significant difference between the groups (p ≤ 0.001), and about half of them were mild (51.8%) according to SAPS II score, while only 11.6% were severe, with no significant differences between the groups (p: 0.130). The median length of ICU stay was 10 days with significant difference between the groups (p ≤ 0.001). The most common respiratory support was HFNC (51.6%); however, 46.9% of the patients have NIV, and only 9.3% of them have IMV with significant difference between the groups (p ≤ 0.001).

Figure 2 shows the characteristics of the readmitted cases, where COPD exacerbation was the most common cause of readmission (35.7%) followed by pneumonia (21.4%) and OSA exacerbation (20.2%) and IPF exacerbation (6%). Most readmitted patients (61.9%) improved, 20.2% of them were discharged on home CPAP, 4.8% on home oxygen, while 10.7% died. Most of the patients were readmitted once (77.4%), while 8.3% were readmitted three times.

Fig. 2

Charactersitics of the cases readmitted to the RICU

Table 2 shows the results of univariate and multivariate binary logistic regression of the risk factors associated with readmission. It showed that age and SAPS II grade were not significantly associated with readmission. Results showed that the significant independent predictors of readmission were being male [AOR (95% CI): 1.8 (1.01–3.1), p: 0.046], having organ failure (renal/liver) [AOR (95% CI): 29.9 (7.9–113.4), p ≤ 0.001], and length of ICU stay more than 12 days [AOR (95% CI): 4.8 (2.6–8.6), p ≤ 0.001]. Also, the discharge status was a significant predictor of readmission, where AOR (95% CI) were 22.03 (6.04–80.3) for being discharged on home oxygen, 8.5 (3.8–18.9) for being discharged on home CPAP, and 9.5 (4.4–20.4) for being shift to ward, with p ≤ 0.001 for each of them. Results also showed that while the type of respiratory failure and the type of respiratory support received were significantly associated with readmission in the univariate analysis, they were insignificant in the multivariate analysis.

Table 2 Univariate and multivariate regression of the risk factors of readmission among patients admitted to the RICU

Table 3 shows the results of univariate and multivariate binary logistic regression of the risk factors associated with inhospital mortality. The most important independent predictor of mortality was receiving IMV [AOR (95% CI): 913.5 (227.5–3667), p ≤ 0.001]. Also, the presence of other organ failure was a significant predictor of mortality [AOR (95% CI): 13.9 (2.6–76.1), p: 0.002]. Although having higher age, higher SAPS II score, longer ICU stay, and type 2 respiratory failure and having COPD, cancer, and HTN were significantly associated with higher mortality in the univariate analysis, none of them was significant in the multivariate analysis.

Table 3 Univariate and multivariate regression of the risk factors of inhospital mortality among patients admitted to the RICU

Discussion

The current study was done in a single specialized ICU which is dedicated only to receiving critically ill patients with respiratory diseases. The readmission rate in our Mansoura University’s RICU during the year 2023 was 10.3%, and COPD exacerbation was the most common cause of readmission (35.7%) followed by pneumonia (21.4%), OSA exacerbation (20.2%), and IPF exacerbation (6%). The readmission rate in our study was slightly lower than one population-based study in Taiwan [9] that reported a readmission rate of 13.13%. This difference may be attributed to the design of the study design that involved multiple medical and surgical ICUs all over the country. According to this study, the most frequent diagnoses of readmissions were as follows: ischemic heart disease (18.49%), followed by diseases of the lung [respiratory failure (10.94%); pneumonia (9.26%)], cerebrovascular disease (5.69%), and sepsis (5.45%).

Another 2-year single-center (medical-surgical ICU) cohort study in Brazil [2] reported that the readmission rate is 10%. Nearly one-third of readmissions (29.3%) to the ICU occurred. Also, another 3-year single medical center study in Korea [10] reported a similar readmission rate is (9.6%). This study included all medical patients admitted to the ICU, not only the cases of respiratory failure. The most common cause of ICU readmission was respiratory failure (87.9%). Nearly one-third of cases (29.3%) were readmitted to the ICU within 72 h or less of discharge, and two-thirds of readmissions occurred after 72 h [2].

A much lower rate of readmission was reported by another 7-year study that was done in Taiwan [11]. It revealed that the readmission rate of mechanically ventilated patients within 7 days after discharge from ICU is 4.5%. Different rates of readmission can be attributed to the difference in the designs of the studies, where studies count the readmission after different durations following discharge. Also, some studies included medical patients, while others include medical and surgical. Our study was done in the respiratory ICU, while other studies included general ICU.

Regarding the risk factors of readmission, our results show that age and SAPII score were not associated with readmission. About half of the enrolled patients were mild (51.8%) according to SAPS II score, while only 11.6% were severe, with no significant differences between the groups (p: 0.130). On the other hand, sex was significantly associated with readmission. Males were nearly two times more likely to be readmitted [AOR (95% CI): 1.8 (1.01–3.1), p = 0.046].

Results reported by previous studies were consistent regarding the association of sex with the readmission rate, even though these studies were different in the design and methodology of the study. In a 1-year study in Turkey [12], it was determined that male patients had higher risk of readmission (p < 0.0010). The study in Brazil [2] found that the patients readmitted to the ICU were more frequently men ([60.6%] vs. [56.1%]; p = 0.042). Another study in Korean found that male sex is independent risk factors [odds ratio (OR) 3.17, 95% confidence interval (CI) 1.29–8.48] [10].

Ponzoni et al. found a significant association between readmission rate and the age which was reported as a significant independent predictor (OR, 0.988; 95% CI, 0.982–0.994; p = 0.001). Also, they found that the higher severity of illness at index ICU admission according to SAPS III score was a significant independent predictor (50 [41–61] vs. 42 [32–54], respectively, for readmitted and non-readmitted patients; p = 0.001) [2]. Comparable results were reported by Jo et al.’ study who found that older age groups were ≥ 75-year-old group (OR = 1.66), and the male individuals showed a higher risk of ICU readmission than the female individuals.

Length of stay is consistently reported to be a significant risk factor of the readmission, whatever the design or methods of the study. According to our results, the median length of ICU stay was 10 days with significant difference between the groups (p ≤ 0.001). Patients who stayed more than 12 days in the ICU were more than four times more likely to be readmitted, even after controlling for confounders [AOR (95% CI): 4.8 (2.6–8.6), p ≤ 0.001]. Ertan et al. reported that cases with a longer ICU stay (> 10 days) had a higher risk for readmission (p: 0.003). Chu et al. reported that the risk of readmission was the highest when the cases stayed in ICU for > 21 days (odds ratio [OR] = 1.5) [11]. Comparable results were also reported by Ponzoni et al.’ study. They found that the length of ICU stay was higher among the readmitted patients compared to the non-readmitted patients, even after adjusting for confounders (p = 0.001) [2].

The most common respiratory support was HFNC (51.6%); however, 46.9% of the patients have NIV, and only 9.3% of them have IMV with significant difference between the groups (p ≤ 0.001). Although receiving respiratory support as IMV significantly increased the likelihood of readmission in univariate analysis, it was insignificant in the multivariate regression. Ponzoni et al. found that the respiratory support needed during ICU admission, such as IMV, noninvasive ventilation, and vasopressors, was independent significant predictor of readmission [use of vasopressors (OR, 1.391; 95% CI, 1.130–1.713; p = 0.002) and length of ICU stay (p = 0.001)] [2].

Type 2 respiratory failure increased the likelihood of readmission more than two times compared to type 1; this association was insignificant after multivariate regression. Elshafey and Hewidy (2014) found that the rate of readmission among patients with type 2 respiratory failure was two times than rate among patients with type 1 respiratory failure [3]. The discharge status of the patients was significantly associated with readmission even after controlling of confounders. For example, patients who were discharged on home oxygen were 22 times more likely to be readmitted compared to patients who improved.

Having comorbidities like cancer and renal/liver failure significantly increase the likelihood of readmission, but in multivariate analysis, only organ failure significantly increased the likelihood of readmission 30 times [AOR (95% CI): 29.9 (7.9–113.4), p ≤ 0.001]. Comparable results were reported by Ponzoni et al. who reported that the presence of comorbidities as chronic kidney disease and liver cirrhosis was a significant independent predictor of readmission as both of them increased the likelihood of readmission two times (p: 0.001). Also, according to their study, cancer and congestive heart failure were significant independent predictors [cancer (OR, 1.4; 95% CI, 1.1–1.7; p = 0.002), congestive heart failure (OR, 1.8; 95% CI, 1.5–2.4; p = 0.005)]. Regarding Jo et al.’s study [10], the presence of comorbidities as diabetes mellitus (OR 3, 95% CI 1.3–7.1) increased the risk of readmission even after controlling for other variables, while other factors either were insignificant as the presence of continuous renal replacement therapy during stay in the ICU (OR 2.8, 95% CI 0.9–9.1) or were significant but with subtle increase in the risk of readmission as WBC on the day of extubation (OR 1.1, 95% CI 1.07–1.2) and heart rate just before ICU discharge (OR 1.03, 95% CI 1.01–1.06).

In the current study, the most important independent predictor of mortality was receiving IMV [AOR (95% CI): 913.5 (227.5–3667), p ≤ 0.001], and presence of other organ failure was a significant predictor of mortality [AOR (95% CI): 13.9 (2.6–76.1), p = 0.002]. Although the patients having higher age, higher SAPS II score, longer ICU stay, type 2 respiratory failure, COPD, cancer, and HTN were significantly associated with higher mortality in the univariate analysis, none of them was significant in the multivariate analysis.

The study conducted in Brazil [2] revealed similar results. They reported that higher SAPS III score [AOR (95% CI): 1.06 (1.05–1.1), p < 0.001], the presence of cancer [AOR (95% CI): 1.3 (1.01–1.6), p = 0.041], use of vasopressors [AOR (95% CI): 1.96 (1.6–2.5), p < 0.001], mechanical ventilation [AOR (95% CI): 2.97 (1.6–2.5, p < 0.001] and renal replacement therapy [AOR (95% CI): 2.1 (1.6–2.7), p < 0.001], and length of ICU stay > 3 days [AOR (95% CI): 1.5 (1.1–1.9), p: 0.004] were also associated with increased risk of inhospital mortality even after controlling for confounders.

Ertan et al. [12] stated that the only significant independent predictors of mortality were age (p: 0.035) and readmission (p: 0.035). Other factors as chronic obstructive pulmonary disease (COPD), APACHE II score, and modified Charlson Comorbidity Index (mCCI) score were significantly associated with mortality in the univariate analysis but not in the multivariate regression analysis.

The results of our study were largely consistent with most of other studies. However, the discrepancies between our study results and other studies results may be attributed to multiple reasons:

• (1) The current study was done in a single specialized ICU dedicated only for critically ill patients with respiratory diseases, while most of other studies include either single general medical ICU or medical-surgical ICU or multicenter ICUs, so the natures of patients diseases and comorbidities widely differ

• (2) Our study was a 1-year study, while most of other studies were done in several years.

Various risk factors were associated with readmission to ICU. According to a recent systematic review that approached the causes of high acute care rehospitalization, three risk factors were associated with rehospitalization 1 year after discharge: the presence of comorbidities, incidents during initial hospitalization (e.g., duration of mechanical ventilation and the presence of delirium), and infection following hospital discharge. Hospital readmission is prevalent among survivors of critical illness. Careful attention to the proper treatment of associated comorbidities during transitions of care may help to decrease healthcare utilization after critical care discharge [13].

Finally, according to the results of our study and most of other comparative studies, during initial admission, it is important to identify a very high-risk subgroup of patients at increased risk for ICU readmission aiming to reduce the rate of readmissions with special emphasis on the discharge status of the patients to improve patient safety, quality of care, and saving ICU resources.

Limitations of the study

The study duration included only 1 year. However, the sample size of the study was sufficient, and 1-year incidence could be obtained. The other limitation was that not all patients were followed up for the same duration.

Conclusion

The rate of readmission was not high (only 10.3%). The most common causes of readmission were COPD exacerbation, followed by pneumonia and OSA exacerbation, respectively. Type of respiratory failure, male gender, the presence of other organ failure, and length of ICU stay are significant independent predictors of readmission in Mansoura University’s RICU during the year 2023. Therefore, higher-risk individuals should receive attention and careful assessment before discharge from the ICU to reduce the rate of readmission.