The safety of morphine in patients with acute heart failure: A systematic review and meta‐analysis

Abstract While morphine has long been widely used in treating acute heart failure (AHF) due to its vasodilatory properties and anticipated anxiolysis, it remains unclear whether the application of morphine to those patients is reasonable. We aim to conduct a systematic review and meta‐analysis to assess the safety of morphine in patients with AHF. We searched PubMed, Cochrane Library, and Embase electronic databases from inception through March 2020. Pooled odds ratios (ORs) and 95% confidence intervals (CIs) were used to estimate the outcomes. Seven studies with 172, 226 patients were included. The results showed that morphine usage was not associated with increased in‐hospital mortality (OR: 1.94; 95% CI 0.93 to 4.03; p = 0.08). However, the use of morphine significantly increased the risk of invasive ventilation (OR: 2.72; 95% CI 1.09 to 6.80; p = 0.03). Furthermore, the subgroup analysis indicated that the application of morphine was not associated with increased 7‐day all‐cause mortality in patients with AHF (OR: 1.69; 95% CI 0.80 to 3.22; p = 0.11) but significantly increased the risk of 30‐day all‐cause mortality (OR: 1.59; 95% CI 1.16 to 2.17; p = 0.004). Based on current evidence, our results suggested that although morphine therapy did not significantly increase the risk of short‐term death (in the hospital or within 7 days) in patients with AHF, the risk of long‐term death and invasive ventilation were significantly increased. This result needs to be further confirmed by an ongoing randomized control trial.


| INTRODUCTION
Acute heart failure (AHF) is a rapidly increasing worldwide health problem with a high mortality rate and a high economic burden that has a high impact on health systems. It is broadly defined as a rapid onset of new or worsening symptoms of heart failure, which itself is not a single disease entity but rather a syndrome of the worsening of signs and symptoms reflecting an inability of the heart to pump blood at a rate commensurate to the needs of the body, 1 requiring urgent hospital admission for evaluation and treatment. 2 Morphine, first isolated in the early 1800s, 3 has been used in patients with AHF due to its anticipated anxiolytic and vasodilatory properties for many years 15 However, to date, the clinical effects of using morphine in patients with AHF remain controversial. It seems that the results from several retrospective studies do not support the safety of morphine in the treatment of patients with AHF. 4,5 The collected evidence suggested that applying morphine to patients with AHF may increase adverse events, which included a greater frequency of mechanical ventilation, prolonged hospitalization, more ICU admissions, and increased in-hospital mortality. However, other studies did not find this association. 6,7 Thereafter, several articles have reviewed published studies but have not given clear support or negative answers about the effects of morphine on patients with AHF. [8][9][10] Despite increasing interest in the field of palliative care for heart failure, to date, there has been no clear uniform standard for the use of morphine. The 2017 Association/American College of Cardiology Heart Failure guidelines 11 do not mention morphine either in the routine treatment of AHF or in late palliative care. In addition, the European Society of Cardiology 12 does not recommend routine use, only for refractory dyspnea as palliative care in advanced HF patients, and suggests that morphine be cautiously considered in patients with severe dyspnea, mostly with pulmonary edema (class IIb recommendation; the level of evidence B) due to side effects from its use, including nausea, hypotension, bradycardia, and respiratory depression (potentially increasing the need for invasive ventilation) in patients with AHF. Thereafter, several articles 13,14 advised against the use of morphine to treat patients with AHF. However, to the best of our knowledge, despite a certain amount of data in the literature, there are no meta-analysis studies that have evaluated the real effects of morphine on patients with AHF.
More importantly, in addition to the anticipated anxiolytic and vasodilatory properties, morphine also showed significant side effects, such as vomiting and aspiration, which may increase mortality. 15,16 Although respiratory depressive effects can be reversed by an appropriate dose of naloxone, morphine-induced hypotension has the potential to decrease myocardial perfusion and increase myocardial ischemia, ultimately resulting in death from cardiogenic shock. It is important, therefore, to comprehensively evaluate the effect of morphine in the treatment of heart failure. 17,18 Therefore, to fully evaluate the benefits and risks in the absence of evidence from any randomized controlled trial (RTC), we conducted a comprehensive meta-analysis of all available studies to elucidate the effect of morphine on patients with AHF.

| Literature search
This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. 19 We systematically conducted a computerized search through the PubMed, Cochrane Library, and EMBASE databases for eligible studies up to March 2020. No geographic restriction was applied in the search process. A comprehensive search strategy was developed based on the following terms: (1) Morphine and intravenous morphine and (2) AHF, acute decompensated heart failure, and acute pulmonary edema. In addition, they were combined with Boolean operators "AND".
Studies were taken into account when they satisfied the following inclusion criteria: (1) Patients treated with morphine as the case group and patients without morphine as the control group; (2) Contained extractable data, such as all-cause mortality, in-hospital mortality, and the risk of invasive ventilation. Exclusion criteria were as follows: (1) Studies not pertinent to morphine or AHF; (2) Lack of a control group; (3) Publication with insufficient data (e.g., abstract, editorial, and comment); and (4) Non-English studies.

| Data extraction and quality assessment
All papers were independently screened by two reviewers (Dandan Zhang and Wei Lai) according to the search strategy. The first phase of screening was performed by reading abstracts, and the second phase of screening involved reviewing the full text. Ultimately, articles meeting the eligibility criteria were further reviewed. Extracted databases were then cross-checked between the two authors to rule out any discrepancies. Additionally, disagreement was resolved by discussion with a third investigator (Kui Hong). If both unadjusted and adjusted ORs existed in one study, we extracted the most completely adjusted ORs. From each study, extracted information included the following elements: name of the first author, year of publication, country, study design, inclusion criteria, total number of patients, proportion of male patients, age, all-cause mortality, in-hospital mortality, and invasive mechanical ventilation. All studies were evaluated using the Newcastle-Ottawa Scale (NOS) for observational studies. The validated NOS items with a total of nine stars involved three terms, including the selection of the population, the comparability of the study, and the assessment of the outcome.

| Statistical analysis
All statistical analyses were performed by using RevMan Manager version 5.3 (The Cochrane Collaboration 2014; Nordic Cochrane Center Copenhagen, Denmark). We used the odds ratio (OR) with 95% confidence intervals (CIs) to evaluate the endpoints. We used the Newcastle-Ottawa quality assessment scale (NOS) to evaluate the quality of all included studies. 20 A study with a NOS score of ≥6 stars was regarded as highquality; otherwise, it was regarded as a low-quality study. 21,22 The choice between fixed or random effects models was determined by evaluating the heterogeneity as recommended by the scientific statement of the American Heart Association. The between-study heterogeneity was assessed quantitatively using the I 2 test. I 2 values of 25%, 50%, and 75% were considered to represent low, moderate, and large heterogeneity, respectively. When significant heterogeneity existed across the included studies, a random effects model was used for the analysis. If this was not the case, the fixed-effect model was used. The statistical significance threshold was set at p < 0.05.

| Characteristics of included studies and quality assessment
Seven studies with 22 967 cases and 172 226 patients with AHF were included in this meta-analysis. [4][5][6][7]14,23,24 Overall, these seven studies were published from 1999 to 2019 and emerged from different countries, including Spain, 14,23 USA, 4,5 Israel, 6,24 and UK. 7 The sample sizes of the included studies varied from 181 to 147 362.
The mean age ranged from 73 to 81 years. The duration of follow-up across the studies varied from 48 h to 30 days. Among the seven studies, most were retrospective case-control studies. The detailed characteristics of the selected studies are presented in Table 1. The methodological quality of the included studies was considered to have a low risk of bias, with 6-7 stars (Table S1).

| In-hospital mortality
Five studies with 22 338 cases/170993 patients contributed to data for this outcome 5,6,14,23,24 The pooled OR showed that there were no differences in in-hospital mortality between the morphine and control groups with high heterogeneity (OR: 1.94; 95% CI 0.93 to 4.03; p = 0.08, I 2 = 96%) ( Figure 2).

| The risk of invasive ventilation
Four studies with 22 047 cases/167847 patients assessed the risk of invasive mechanical ventilation. 4,5,14,24 The results showed that morphine treatment was associated with an increased risk of invasive ventilation incidence (OR 2.72; 95% CI 1.09 to 6.80; p = 0.03, I 2 = 93%) in patients with AHF compared with the control group ( Figure 3).

| 7-day and 30-day all-cause mortality
Three studies with 1175 cases/9904 patients reported the association between morphine therapy and all-cause mortality in patients with AHF. 6,7,14 Subgroup analysis of all-cause mortality was performed according to different follow-up times between the morphine and control groups. As shown in Figure 4, the summary OR showed that there was no difference in 7-day all-cause mortality (OR: 1.69; 95% CI 0.89 to 3.22; p = 0.11, I 2 = 61%) between the morphine and control groups. However, morphine therapy was associated with significant 30-day all-cause mortality (OR: 1.59; 95% CI 1.16 to 2.17; p = 0.004, I 2 = 0%) with no evidence of heterogeneity.

| Publication bias
Publication bias was not performed in this study, as the publication bias could not be ascertained, as the number of studies included for each item was <10. 25

| DISCUSSION
To the best of our knowledge, this is the first meta-analysis to evaluate the clinical outcomes of morphine in patients with AHF. The significant findings can be summarized as follows: (1) there was no difference in in-hospital mortality or 7-day death between the morphine and control groups in patients with AHF; (2) morphine was associated with a higher risk of invasive ventilation and long-term allcause mortality in patients with AHF.
Although morphine is used for patients with AHF based on its beneficial effects, including decreasing the preload and afterload of the heart and improving anxiety, respiratory difficulty, and chest pain, 26  Moreover, we also studied the role of the follow-up time in the present meta-analysis. We found that morphine was associated with 30-day all-cause mortality but not short-term death (7-day all-cause mortality and in-hospital mortality). This result should be treated with caution. The reason that morphine did not increase short-term mortality might be that patients with AHF received multidrug therapy to relieve AHF during the 7 days after admission, which obscured the side effects of morphine to some extent. Moreover, comorbidities and drug interactions had a greater impact on short-term mortality than on long-term mortality, which could not reflect exactly the real pharmacological effect. Furthermore, we carefully analyzed the two articles, which included long-term mortality. we speculated that the possible cause includes the prolonged adverse cardiovascular effect, modulation of receptor sensitivity, respiratory drive inhibition, or chronic negative inotropic action. However, only two studies were included in each subgroup analysis, which might bias the summary results. Given the above, the 7-day all-cause mortality might be close to in-hospital mortality to some extent, and thus extending the follow-up time of patients with AHF is essential.
Another point worth discussing is that the safety of morphine in patients with AHF was still controversial in different studies after propensity score matching. Oscar et al. concluded that morphine increased all-cause mortality at different times in the EAHFE registry analysis, which conflicted with negative results given by the Israeli HFSIS registry 6 and 3CPO trial. 7 This might occur because the former analyzed all 46 baseline study variables, and the propensity score was ultimately calculated using 24 significant baseline variables, which was significantly more than the number of variable factors in the latter two studies. In addition, the number of patients included in the previous study was much greater. Therefore, the propensity score method cannot substitute for a prospective RTC. Future RCTs should be conducted to observe whether patients with AHF should receive morphine therapy.
However, these risks do not mean that morphine cannot improve patients' symptoms and subjective perception. As mentioned earlier, the mechanism by which morphine improves AHF symptoms includes inhibition of sympathetic nerve activity, sedative and anxiolytic effects, and improvement of lung ventilation. Therefore, it is more plausible that potential pharmacological effects overshadow the benefits for patients with AHF. The aim is to comprehensively assess the balance between risks and improvement of the symptoms through long-term follow-up. It is more reasonable for us to evaluate the effect of morphine for patients with AHF more comprehensively in this way.
Of course, we must point out again that our results are consistent with the current guideline recommendations, categorized as IIb or at the level of evidence (B). The evidence collected likely would not suffice to support the safety of morphine in the treatment of patients with AHF. The latest studies found that both the deterioration of renal function and the incidence of cardiogenic shock in patients with AHF are related to the use of morphine. 28,29 Undeniably, more RCTs should be conducted to observe whether patients with AHF should receive morphine. It is only with RCTs with adequate sample sizes that we will be able to provide a reliable answer to the question as to whether the use of morphine is necessary. Consequently, indications of the Guidelines of the European Society of Cardiology 12 and the American Heart Association/American College of Cardiology 11 are likely to stand until more extensive scientific data becomes available.
To date, no RCTs have assessed the effect of morphine on inhospital mortality or all-cause mortality in patients with AHF. However, we noted that there is one ongoing multicenter and randomized control trial (MIMO trial) that aims to assess the efficacy and safety of morphine, 23 which will not only fill our gaps in knowledge on the adverse effects and risks associated with morphine but also help to guide clinical F I G U R E 1 Flowchart of the study selection process for the meta-analysis decisions regarding the use of morphine in patients with AHF. Midazolam not only has been shown to be an efficient anxiolytic but also has positive cardiovascular effects. 30 More specifically, compared with morphine, midazolam has sedative effects by allosterically increasing the affinity of GABAA receptors for GABA without serious side effects such as nausea, vomiting, and hypotension. 15 Therefore, more importantly, the safety and effectiveness of another drug, midazolam, will be evaluated, which may be a viable alternative treatment. 23

| Study limitations
The quality of the studies included in this review was acceptable according to the quality assessment. However, the present meta-analysis has several study limitations. First, the number included in this meta-analysis was limited, and all the studies included were observational studies, which cannot adequately prove the causation between morphine and high mortality and the incidence of mechanical ventilation. The findings, therefore, probably have been influenced by selection bias or another residual confounding factor. Despite studies attempting to use propensity score matching to minimize selection bias, residual confounding can never be fully excluded. Performing regression analyses to account for these confounders. However, meta-regression is no recommended by the guideline when the included studies are limited (N < 10). Therefore, we did not perform a meta-regression in the present study.
It is worth noting that propensity score matching inevitably greatly reduced the sample size. As Caspi et al. 31  Second, most studies did not provide a specific description of morphine therapy, such as drug dose and administration route.
Therefore, due to data restrictions, we could not further analyze the effects of effective dosage or the different administration approaches of morphine in patients with AHF. Only Caspi O et al. 24 investigated the effects of administered amounts and found that there was a significant linear dependency between the incidence of invasive mechanical ventilation and the dose of morphine.
Third, the follow-up time of the included studies was short. Last, it is difficult to evaluate the potential pharmacological effect in vasodilatation and dyspnea amelioration of morphine owing to a lack of data regarding the degree of difficulty in breathing and changes in blood pressure. In summary, the comprehensive effects of morphine on patients with AHF need to be confirmed by more clinical information.

| CONCLUSIONS
Morphine application could be associated with the increased incidence of invasive ventilation and 30-day all-cause mortality in patients with AHF. Future RCTs are warranted to further assess whether patients with AHF should receive morphine therapy.

CONFLICT OF INTEREST
All the authors declared that they have no conflicts of interest.

DATA AVAILABILITY STATEMENT
All data generated or analyzed during this study are included in this published article.