Efficacy of Non-invasive Ventilation Compared with Invasive Mechanical Ventilation in Cardiogenic Shock: A Systematic Review and Meta-analysis

Authors: Tanawat Attachaipanich, Kotchakorn Kaewboot, Suthinee Attachaipanich

Abstract

Background

Cardiogenic shock has a high mortality rate, and ventilation support is commonly required. Non-invasive ventilation (NIV) provides several advantages over invasive mechanical ventilation (IMV) and is safe in cardiogenic pulmonary edema. However, its efficacy in cardiogenic shock remains unclear. This study aimed to evaluate the efficacy and safety of NIV compared to IMV in this population.

Data sources

A systematic search was conducted across four databases, including PubMed, Embase, Web of Science, and Cochrane CENTRAL, from inception to February 12, 2025, without language restrictions.

Study selection

Studies were included if they compared NIV vs IMV in patients diagnosed with cardiogenic shock and reported relevant clinical outcomes.

Data extraction

Two reviewers independently extracted data on study characteristics, patient demographics, and outcomes. Risk of bias was independently assessed using appropriate tools for the included studies.

Data synthesis

A total of 6 observational, non-randomized studies involving 2,302 participants were included in this meta-analysis, using a random-effects model. Non-invasive ventilation was associated with a significantly lower risk of in-hospital mortality [odds ratio (OR) 0.55 (95% confidence interval (CI): 0.38–0.79), p < 0.01; I^2^ = 35.85%] and 30-day mortality (OR: 0.47 (95% CI: 0.35–0.63), p < 0.01; I^2^= 0%) compared to IMV. Non-invasive ventilation also reduced the length of hospital stays and intensive care unit/coronary care unit stays. Sensitivity analyses, including two studies using propensity score adjustment, showed the consistent benefit of NIV in reducing in-hospital mortality and length of hospital stay.

Conclusions

Non-invasive ventilation was associated with improved short-term outcomes compared with IMV in carefully selected patients with cardiogenic shock. However, all included studies were observational, with potential for residual confounding; therefore, further randomized studies are warranted.

How to cite this article

Attachaipanich T, Kaewboot K, Attachaipanich S. Efficacy of Non-invasive Ventilation Compared with Invasive Mechanical Ventilation in Cardiogenic Shock: A Systematic Review and Meta-analysis. Indian J Crit Care Med 2026;30(2):148–154.

Highlights

• Non-invasive ventilation was associated with lower in-hospital and 30-day mortality, without a significant difference in 60-day mortality, compared with IMV in cardiogenic shock.
• Non-invasive ventilation shortened hospital and ICU/CCU stays.
• Benefits were consistent in propensity score-adjusted analyses.
• Non-invasive ventilation may be most effective in early-stage shock; randomized trials are still needed.

Introduction

Cardiogenic shock is defined as a state of low cardiac output resulting in end-organ hypoperfusion and hypoxia.^1^ The most common cause of cardiogenic shock is acute myocardial infarction (AMI).^1^ Despite a decline in the incidence of ST-elevation myocardial infarction, the incidence of cardiogenic shock among AMI patients has remained stable at 8–10%.^2^ Even with advancements in reperfusion therapy and mechanical circulatory support, in-hospital mortality remains high, ranging from 30 to 50%.^3^ One randomized trial reported that mortality in AMI-associated cardiogenic shock exceeds 50%.^4^

The hemodynamic changes of cardiogenic shock are primarily associated with increased pulmonary capillary wedge pressure, leading to cardiogenic pulmonary edema and the need for ventilatory support. Previous studies reported that up to 80% of patients with cardiogenic shock require invasive mechanical ventilation (IMV).^5^ Current guidelines recommend IMV in the setting of cardiogenic shock; however, the supporting evidence is limited.^1,6,7^

Non-invasive ventilation (NIV) has been shown to be effective in patients with cardiogenic pulmonary edema and has seen increased utilization due to several advantages over IMV, including a reduced risk of ventilator-associated complications. In patients with cardiogenic pulmonary edema without cardiogenic shock, NIV has been associated with a reduced need for IMV and lower in-hospital mortality.^8^ However, extrapolation of these benefits to patients with cardiogenic shock remains uncertain. In this population, NIV may worsen hypotension by reducing preload and cardiac output, may delay necessary endotracheal intubation, and may be poorly tolerated in patients with altered mental status or respiratory muscle fatigue, which are commonly encountered in advanced shock.

Importantly, existing systematic reviews and meta-analyses have not focused exclusively on patients with cardiogenic shock, instead combining populations with cardiogenic pulmonary edema and shock, thereby limiting applicability to this high-risk subgroup. Moreover, recent large contemporary observational registries, including FRENSHOCK and AltShock-2, have provided substantial new data on ventilatory strategies in cardiogenic shock that have not been previously synthesized. To address this knowledge gap, we conducted a systematic review and meta-analysis to evaluate the efficacy and safety of NIV compared with IMV specifically in patients with cardiogenic shock.

Methods

Search Strategy, Selection Criteria, and Data Extraction

A systematic review and meta-analysis were conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.^9^ The study protocol was registered with the PROSPERO International Prospective Systematic Evaluation Registry (CRD420251047511). A systematic search was performed across four databases, including PubMed, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials (CENTRAL), from inception to February 12, 2025, without language restrictions. Non-English abstracts were initially screened using English abstracts when available or were translated using standardized translation tools. Search terms related to cardiogenic shock and NIV were used, with full search strategies detailed in the Supplementary Data 1.

This systematic review and meta-analysis were designed according to the PICOs framework. The population included adult patients with cardiogenic shock (P). The intervention was NIV (I), compared with IMV (C). The primary outcome was in-hospital mortality, and secondary outcomes included 30-day and 60-day mortality, length of hospital stay, and length of intensive care unit (ICU)/coronary care unit (CCU) stay (O). The inclusion criteria for studies in this systematic review and meta-analysis were as (1) studies enrolling patients with cardiogenic shock, (2) studies comparing NIV with IMV, (3) studies reporting all-cause mortality, and (4) studies designed as randomized trials or observational studies. Studies were excluded if they were case series, case reports, conference abstracts, or small studies with fewer than five participants in each treatment arm to minimize small-study bias and unstable effect estimates. Studies enrolling mixed shock populations were included only if outcomes specific to patients with cardiogenic shock could be clearly identified and extracted. Studies in which cardiogenic shock-specific data could not be separated from other shock etiologies were excluded.

Two investigators (T.A. and K.K.) independently reviewed abstracts and full texts to select eligible studies. Discrepancies were resolved through discussion and reviewed by three reviewers (T.A., K.K., and S.A.). A final consensus was made with agreement among all three reviewers. The primary outcome was in-hospital mortality. Secondary outcomes were 30-day mortality, 60-day mortality, length of hospital stay, and length of ICU/CCU stay.

The data extraction from eligible studies was independently conducted by two investigators (T.A. and K.K.) using a standardized data extraction sheet. Extracted data included study design, country, study population, comorbidities, etiology of cardiogenic shock, vasopressor, shock stage, and funding sources. In cases where multiple studies included the same or overlapping participants, only the study with the largest sample size was included in the analysis. When reported, escalation from NIV to IMV was recorded as NIV failure. Handling of NIV-to-IMV crossover varied across studies, followed the original study definitions, and is summarized in Supplementary Table 1.

Quality Assessment

The methodological quality of randomized studies was independently evaluated by two reviewers (T.A. and K.K.) using the Revised Cochrane risk-of-bias tool for randomized trials (RoB 2).^10^ Non-randomized studies were assessed by two reviewers (T.A. and K.K.) using the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I), which evaluates bias across seven confounding, participant selection, classification of interventions, deviations from intended interventions, missing data, outcome measurement, and selection of reported results.^11^ Discrepancies were resolved through discussion and reviewed by three reviewers (T.A., K.K., and S.A.).

Data Synthesis Methods

A pairwise meta-analysis was performed to synthesize the data in this study. Statistical heterogeneity was assessed using the I^2^ statistic and the Cochran's Q test. The Restricted maximum-likelihood model was used to estimate the pooled effect with a 95% confidence interval (CI). Sensitivity analyses were performed, including only studies using propensity score adjustment. For the differences in length of hospital stay or ICU/CCU stay, we pooled estimates from all studies using the weighted mean difference (WMD). Publication bias was evaluated using a funnel plot and Egger's test.

A p-value of less than 0.05 was considered statistically significant. Statistical analyses were conducted using STATA version 16.1 software (StataCorp LLC, College Station, USA).

Results

Characteristics of Included Studies

After excluding duplicate results, the literature search yielded a total of 665 studies. Following title and abstract screening, 615 studies were excluded. A total of 50 studies underwent full-text review. Finally, six studies were included in this meta-analysis, comprising 2,302 participants.^12–17^ A diagram summarizing the flow of study selection is presented in Figure 1.^18^

The included studies were published between 2011 and 2024. The number of participants in these studies ranged from 163 to 503. All included studies were non-randomized, and two studies conducted propensity score adjustment.^12,13^ The mean age of participants ranged from 61 to 74 years. Most of the participants have cardiogenic shock due to acute coronary syndrome. The study characteristics are summarized in Table 1.

Risk of Bias

Six non-randomized studies were considered to have a serious risk of overall bias, according to ROBIN-I.^12–17^ These studies were primarily downgraded due to a serious risk of bias from confounding and classification of interventions. The summarized risks of bias are presented in Supplementary Figure 1.

In-hospital Mortality

Five studies involving 1,799 participants were included in the meta-analysis of in-hospital mortality.^12,13,15–17^ Non-invasive ventilation use was associated with a lower risk of in-hospital mortality compared to IMV, with an odds ratio (OR) of 0.55 (95% CI: 0.38–0.79), p < 0.01. The forest plot for in-hospital mortality is shown in Figure 2. Cochran's Q test showed no statistical significance, and I^2^ was 35.85%, indicating low heterogeneity. The funnel plot is presented in Supplementary Figure 2, and the Egger test was not statistically significant, indicating no evidence of publication bias.

Sensitivity analysis, which included only studies using propensity score adjustment, demonstrated the association between NIV use and a lower risk of in-hospital mortality, with OR 0.48 (95% CI: 0.26–0.91). The forest plot for subgroup analysis of in-hospital mortality is represented in Supplementary Figure 3.

30-day Mortality

The pooled results from 4 studies involving 1,534 participants showed that NIV use was associated with a lower risk of 30-day all-cause mortality compared to IMV, with an OR of 0.47 (95% CI: 0.35–0.63), p < 0.01.^13–15,17^ The forest plot for 3-month mortality is shown in Figure 3. Cochran's Q test showed no statistical significance, and I^2^ was 0%, indicating low heterogeneity.

60-day Mortality

The pooled results from 2 studies involving 622 participants showed that there was no difference in 60-day mortality between NIV and IMV, with an OR of 0.62 (95% CI: 0.38–1.02), p = 0.06.^13,15^ The forest plot for 6-month mortality is shown in Supplementary Figure 4. Cochran's Q test showed no statistical significance, and I^2^ was 18.62%, indicating low heterogeneity.

Length of Hospital Stay

There were three studies involving 708 participants included in the meta-analysis of length of hospital stay.^12,13,17^ Non-invasive ventilation use was associated with lower length of hospital stay compared to IMV, with a WMD of -3.20 (95% CI: −5.33 to −1.07), p < 0.01. The forest plot for the length of hospital stay is shown in Supplementary Figure 5. Cochran's Q test was non-significant, and I^2^ was 0%, indicating low heterogeneity.

Sensitivity analysis, which included only studies using propensity score adjustment, also showed that NIV use was associated with lower length of hospital stay compared to IMV, with a WMD of -2.62 (95% CI: −5.03 to −0.21), p = 0.03, as shown in Supplementary Figure 6.

Length of ICU/CCU Stay

There were three studies involving 735 participants included in the meta-analysis of length of ICU/CCU stay.^12,13,17^ Non-invasive ventilation use was associated with lower length of ICU/CCU stay compared to IMV, with a WMD of -2.06 (95% CI: -2.76 to -1.37), p < 0.01. The forest plot for the length of ICU/CCU stay is shown in Supplementary Figure 7. Cochran's Q test was non-significant, and I^2^ was 0%, indicating low heterogeneity.

Sensitivity analysis, which included only studies using propensity score adjustment, also showed that NIV use was associated with lower length of ICU/CCU stay compared to IMV, with a WMD of −1.97 (95% CI: −2.68 to −1.27), p < 0.01, as shown in Supplementary Figure 8.

Discussion

To the best of our knowledge, this is the first meta-analysis evaluating the efficacy and safety of NIV in the setting of cardiogenic shock. Our meta-analysis, which included 6 studies comprising 2,302 participants, demonstrated that NIV use was associated with lower in-hospital mortality and 30-day mortality compared to IMV. However, no significant difference was observed in 60-day mortality between NIV and IMV. Additionally, NIV use was associated with shorter hospital and ICU/CCU stays. Sensitivity analyses limited to studies using propensity score adjustment further supported the association between NIV use and reduced in-hospital mortality, as well as shorter duration of hospital and ICU/CCU stays.

Non-invasive ventilation has been shown to be an effective treatment for acute cardiogenic pulmonary edema. A previous meta-analysis, including 2,916 participants with cardiogenic pulmonary edema, demonstrated that NIV reduces in-hospital mortality.^8^ Respiratory failure is common in cardiogenic shock, with reported incidence rates ranging from 60 to 90%.^19,20^ The positive inspiratory pressure from NIV and IMV improves oxygenation, reduces the work of breathing, and enhances cardiac output by decreasing both preload and afterload in heart failure.^21,22^ Non-invasive ventilation provides the advantage of reducing complications associated with IMV, including ventilator-associated pneumonia, aspiration pneumonia, airway trauma, and the need for sedative medications. Current guidelines recommend NIV in acute heart failure to reduce the need for IMV.^23^ However, no randomized controlled trials have evaluated NIV specifically in patients with cardiogenic shock. Although randomized trials of CPAP and NIV in cardiogenic pulmonary edema have demonstrated clinical benefit, these findings cannot be directly extrapolated to cardiogenic shock.

The utility of NIV in cardiogenic shock may be limited by hypotension and shock, which can lead to impaired consciousness and reduced respiratory muscle perfusion. Positive intrathoracic pressure generated by NIV can reduce venous return and preload, potentially worsening hypotension in patients with impaired cardiac output.^24^ In addition, cardiogenic shock is frequently associated with altered mental status and reduced respiratory drive, which may impair patient tolerance and contribute to high rates of NIV failure. NIV failure and delayed endotracheal intubation are recognized predictors of adverse outcomes in shock and may further offset potential benefits of NIV in this population.^25^ Despite the limited evidence on the efficacy of NIV in cardiogenic shock, real-world data indicate increasing utilization, with reported rates ranging from 5 to 16%.^13,15,16^ Carefully selected patients, particularly those in the early stages of cardiogenic shock, may benefit most from NIV.^1^

Only two of the included studies reported baseline Society of Cardiovascular Angiography and Interventions (SCAI) cardiogenic shock classification. In these studies, approximately half of the participants were classified as SCAI class C in both NIV and IMV groups.^12,15^ A propensity-score matched study reported that, among SCAI class C patients, NIV was associated with lower in-hospital mortality (10.1% in the NIV group vs 32.9% in the IMV group).^12^ However, no significant difference in in-hospital mortality was observed between NIV and IMV in patients classified as SCAI class D or E.^12^ Similarly, another cohort study demonstrated lower 60-day mortality in NIV-treated patients with SCAI class C shock but not in those with class D or E shock.^15^ These findings suggest that NIV may be most beneficial in the early stages of cardiogenic shock. However, inconsistent reporting of SCAI stage across studies represents a major limitation, as shock severity strongly influences NIV failure rates and clinical outcomes, thereby limiting interpretation and generalizability.

The failure rate of NIV, requiring escalation to IMV, ranged from 20% to 50% in the included studies.^12,13,15^ This NIV failure rate is substantially higher than the 1.5% reported in acute heart failure settings.^26^ The variability in NIV failure rates among cardiogenic shock patients is likely attributable to patient heterogeneity. Patients with more severe cardiogenic shock, particularly those classified as SCAI class D or E or those with shock secondary to acute coronary syndrome, have a higher risk of NIV failure.^12^

Patients with cardiogenic shock who received IMV were generally more critically ill and had a greater need for pharmacologic and mechanical circulatory support, as well as coronary revascularization.^15,17^ These baseline imbalances reflect clinical selection of IMV for patients with more severe shock and represent a potential source of confounding. To account for these potential confounders, we conducted a sensitivity analysis, including only studies that used propensity score adjustment. This analysis further supported the association between NIV use and improved clinical outcomes compared to IMV, supporting the robustness of our findings.^12,13^ However, this association should be interpreted cautiously, as it may reflect confounding by indication rather than a causal benefit of NIV. Nevertheless, residual confounding due to unmeasured differences in cardiogenic shock severity cannot be excluded. Importantly, such confounding may influence not only in-hospital mortality but also downstream outcomes, including 30-day mortality, which may be affected by subsequent management decisions such as revascularization strategies and use of mechanical circulatory support.

Early mortality among more critically ill patients receiving IMV may artificially shorten hospital length of stay. However, in our analysis, NIV was associated with a shorter length of stay, which may reflect greater shock severity in the IMV group and the associated need for more intensive support, potentially prolonging ICU stay. Overall, length-of-stay findings should be interpreted cautiously.

There are several limitations to this meta-analysis. Given the nature of cardiogenic shock, conducting randomized controlled trials is challenging. All of the included studies were observational and, therefore, subject to potential confounding. To address this, we performed sensitivity analyses, exclusively including studies that used propensity score adjustment, which attempts to balance baseline characteristics and minimize potential confounding. However, residual unmeasured confounding cannot be entirely excluded. Additionally, the limited number of studies included in this meta-analysis may introduce selection bias and contribute to heterogeneity. Furthermore, because only four to six studies were available for each outcome, assessment of publication bias using funnel plots and Egger's test may be underpowered and should be interpreted with caution.

Conclusion

Our meta-analysis suggests that NIV may be associated with improved short-term outcomes in carefully selected patients with cardiogenic shock, particularly those in the early stage of cardiogenic shock. Non-invasive ventilation use was associated with reduced short-term mortality and a shorter length of hospital stay compared to IMV. However, these findings are based on observational studies and should be interpreted with caution due to potential confounding and selection bias. Further randomized studies or large prospective studies are needed to establish the safety and efficacy of NIV and to better define the patient population most likely to benefit. Given the heterogeneity of cardiogenic shock, future studies should stratify patient subgroups based on shock severity to determine appropriate treatment strategies.

Author Contributions

Tanawat Attachaipanich: Conceptualization; Formal analysis; Methodology; writing – original draft; writing – review and editing. Kotchakorn Kaewboot: Conceptualization; Formal analysis; Methodology; writing – original draft; writing – review and editing. Suthinee Attachaipanich: Conceptualization; Formal analysis; Methodology; writing – original draft; writing – review and editing.

PROSPERO Registry Number

CRD420251047511

Supplementary Materials

The supplementary Tables and Figures are available on its journal website www.ijccm.org

Orcid

Tanawat Attachaipanich https://orcid.org/0000-0002-1514-4061

Kotchakorn Kaewboot https://orcid.org/0009-0004-1492-4551

Suthinee Attachaipanich https://orcid.org/0009-0005-3636-0037