Intervention by clinical pharmacists can improve blood glucose fluctuation in patients with diabetes and acute myocardial infarction: A propensity score‐matched analysis

Abstract Acute phase hyperglycemia and exaggerated glucose fluctuation may be associated with poor outcomes in diabetic patients after acute myocardial infarction (AMI). This study aimed to determine whether intervention by clinical pharmacists can mitigate blood glucose and glucose fluctuations in these fragile patients. This retrospective study enrolled patients with diabetes and AMI, from 1 January 2019 to 30 June 2020 in our institution. Blood glucose and glucose fluctuations were calculated before and after the pharmacist's intervention and between patients who underwent intervention and those who did not. Propensity score matching (PSM) was used to reduce the impact of patient characteristics on the results. A total of 170 patients were included in our primary analysis, including 29 patients who received the pharmacist intervention and 141 patients who did not. After the pharmacist's intervention, blood glucose (fasting blood glucose‐FBG, from 11.9 to 9.8; postprandial blood glucose‐PBG, from 15.3 to 13.2; mean blood glucose‐BG, 14.5 to 12.3 mmol/L; p < .001), and glucose fluctuations (standard deviation of blood glucose‐SDBG, from 3.8 to 3.0, mmol/L, p = .005) were significantly improved. Before PSM, no clear effects were found in intervention versus nonintervention patients, in terms of blood glucose and glucose fluctuation indicators, except for FBG (9.3 vs. 8.0. mmol/L, p = .005). Further analysis indicated a high incidence of FBG <7.8 mmol/L in nonintervention versus intervention patients (51.5% vs. 27.6%, p = .003). After PSM, a significant reduction in blood glucose fluctuation (SDBG, 3.0 vs. 4.1, p = .031; PBGE, 2.1 vs. 4.1, p = .017; LAGE, 4.7 vs. 7.2, mmol/L, p = .004), and PBG (11.1 vs. 13.0, mmol/L, p = .048) was observed in the intervention group than in the nonintervention group. The clinical pharmacist intervention contributed to improved outcomes, specifically, in reducing blood glucose fluctuations and potential hypoglycemia risk.

study enrolled patients with diabetes and AMI, from 1 January 2019 to 30 June 2020 in our institution. Blood glucose and glucose fluctuations were calculated before and after the pharmacist's intervention and between patients who underwent intervention and those who did not. Propensity score matching (PSM) was used to reduce the impact of patient characteristics on the results. A total of 170 patients were included in our primary analysis, including 29 patients who received the pharmacist intervention and 141 patients who did not. After the pharmacist's intervention, blood glucose (fasting blood glucose-FBG, from 11.9 to 9.8; postprandial blood glucose-PBG, from 15.3 to 13.2; mean blood glucose-BG, 14

| INTRODUC TI ON
Diabetes mellitus (DM) is an important and frequent comorbidity in patients with acute myocardial infarction (AMI). DM is being present in nearly 30% of AMI cases. 1 AMI patients with DM have a significantly higher mortality than those without diabetes. 2 Many studies have shown a relationship between high blood glucose levels on admission and an increased risk of mortality and poor outcomes after AMI. Studies further verified the association between glucose fluctuation during the phases of AMI and the extent of myocardial salvage. 3 Rapid blood glucose fluctuation levels increase oxidative stress and are even more detrimental than sustained hyperglycemia.
As a result, it is suggested that interventional trials in type 2 DM (T2DM) should not only focus on the blood glucose but also on acute glucose swings. 4 Glucose fluctuations are also closely related to electrocardiographic surrogate markers of impaired myocardial salvage in AMI after reperfusion therapy, 5 and have been shown to be correlate with endothelial dysfunction and atherosclerotic development. 4 Accordingly, glucose fluctuation has been found to be a negative prognostic factor for patients with AMI, 6 and may be a potential detrimental factor for salvaging ischemic damage. 3 The reduction of glucose fluctuation may offer a potentially therapeutic strategy for decreasing myocardial reperfusion injury in AMI patients. 5 Many components, including beta-cell function, diet, exercise, and drugs, contribute to glucose fluctuations. 7 Specifically, improper use of hypoglycemic drugs, poor adherence to medication by the patients, irregular food and exercise patterns, and nonstandard insulin injection will certainly cause glucose fluctuation. 7 What is already known about this subject: • Acute phase hyperglycemia and exaggerated glucose fluctuations may be associated with poor outcomes in diabetic patients after acute myocardial infarction (AMI).

What this study adds:
• Clinical pharmacist's intervention can mitigate blood glucose and glucose fluctuations in diabetes and acute myocardial infarction (AMI) patients in the hospital.
Based on the glucose management goals of these fragile patients, pharmacist's intervention also may reduce potential hypoglycemia risk.
unavailable. This study included two parts. In the first part, clinical pharmacists carried out interventions for the patients with uncontrolled blood glucose and evaluated changes in the blood glucose and fluctuations after interventions. In the second part, comparison between patients who underwent pharmacist intervention and those who did not was performed using the propensity score matching (PSM) analysis. The study protocol was approved by the ethics committees of Renji Hospital, School of Medicine, Shanghai Jiaotong University (KY2019-076) and informed consent was obtained from each patient or their family members.

| The clinical pharmacist intervention model
Pharmacists working in the endocrinology department are skilled in glucose management of diabetic patients and have more than

| Specific pharmacist interventions for controlling glucose fluctuations
The specific pharmacist interventions included consultation-based health education and drug optimization. First, for each patient, the

Inpatient Area
Application for pharmaceutical consultation

Medical Department
Receive and confirm consultation request

| Data collection
Patient characteristics (demographics, diagnosis, and diabetesrelated indices) were recorded by reviewing the medical charts and hospital information system. The blood glucose levels were detected using Contour TS blood glucose meter (Bayer HealthCare) and Contour TS blood glucose test strips. Finger prick blood glucose data were used to evaluate the blood glucose fluctuations. The blood glucose data including fasting blood glucose (FBG, 6:00 am), postbreakfast blood glucose (PBG, 9:00 am), post-lunch blood glucose (PBG, 13:00 pm), and post-dinner blood glucose (PBG, 19:00 pm) were obtained from the blood glucose test records in the electronic system. In the first part, glucose data were collected 4 days before and after the pharmacist consultation. In the second part, as part of the PSM analysis, the FBG data were collected 3 days before discharge and the PBG data were collected 24 h before discharge.

| Outcomes measures
The

| Glucose levels and glucose fluctuations before and after the pharmacist intervention
After the pharmacist intervention, the glucose indicators showed a significant improvement (mean FBG decreased from 11.9 mmol/L to

| Glucose levels and glucose fluctuations in the intervention and nonintervention patients
Before PSM, as for the glucose indicators, there were no significant differences between the intervention and nonintervention patients in terms of mean PBG (11.7 mmol/L vs. 11.6 mmol/L, p = .841) and mean BG (11.0 mmol/L vs. 10.7 mmol/L, p = .542). Unexpectedly, we found a seemingly opposite result. The mean FBG was higher in the intervention group than in the nonintervention group (9.3 mmol/L vs. 8.0 mmol/L, p = .005) ( Figure 4A). According to the patient's condition and the reasons for admission, the glycemic goals of AMI patients should be set to less stringent levels, for example, the level of AMI patients, FBG 7.8-10 mmol/L and PBG 7.8-13.9 mmol/L, based on Chinese endocrinologist consensus on blood glucose management for Chinese inpatients. 13 As such, FBG <7.8 mmol/L may be considered as a potential risk for hypoglycemia in patients with AMI, even though hypoglycemia did not occur in these patients.
After further analysis, we found a higher incidence of FBG<7.8 mmol/L in the nonintervention patients than in the inter-   Figure 4D).

| DISCUSS ION
Our study focused on a new pharmacist-clinician collaboration model based on the addition of pharmacist consultations to the AMI patients with DM, with the goal of optimizing comprehensive care for these patients. In this study, the pharmacist intervention was associated with improvements in the blood glucose levels (mean F I G U R E 3 Blood glucose and fluctuation data before and after pharmacist consultation. Blood glucose (A) before and after pharmacist consultation, which contains FBG, PBG, and BG. Glucose fluctuation (B) before and after pharmacist consultation, which contain SDBG, PBGE, and LAGE. Blood glucose data collected from blood glucose monitoring from fasting blood glucose (FBG, 6:00 am), post-breakfast blood glucose (PBG,9:00 am), post-lunch blood glucose (PBG,13:00 pm), and post-dinner blood glucose (PBG,19:00 pm). Continuous data were collected 4 days before and 4 days after pharmacist consultation. BG, blood glucose; SDBG, standard deviation of blood glucose; PBGE, postprandial glucose excursion; LAGE, largest amplitude of glycemic excursions. Data were described as mean ± SE, p < .05 was considered as a significant difference. ** p < .01 and * p < .05 FBG, mean PBG, and mean BG) as well as in glucose fluctuations (SDBG, PBGE, and LAGE). Furthermore, the clinical pharmacists were able to reduce the potential risk of hypoglycemia and make the blood glucose control more achievable for these frail patients.
Clinical pharmacists practicing in medical institutions are easily accessible and can play a vital role in providing timely advice for diabetic patients and therapeutic advice for the interdisciplinary care team. 14 The most effective components of the pharmacist intervention were patient-centered services and multidisciplinary care. 15 A German study 16 reported that pharmacists who provided independent case management were more effective when collaborating with clinicians within a multidisciplinary team. In this study, we explored a new pharmacist-clinician collaboration model based on pharmacist consultations, and found that the blood glucose levels and glucose fluctuations of diabetic patients were significantly improved through the pharmacist interventions model.

The rising mortality risk in patients with diabetes during and
after AMI is of critical concern. There is an urgent need for better treatment choices in these patients in addition to intensive medications. 1 Stringent control of risk factors, such as blood glucose, may be a good long-term surveillance strategy in diabetic patients with AMI. 17 The mean BG is an important predictive factor for in-patient mortality, and has been indicated to be independently related to mortality in critically ill patients. 18 A previous umbrella meta-analysis reported that the pharmacist interventions mainly resulted in favorable improvements in HbA1c, FBG, and other cardiovascular risk factors. 11 In our study, we not only assessed FBG, PBG, and mean F I G U R E 4 Blood glucose and fluctuation data before and after propensity score matching (PSM). Blood glucose (A) and fluctuation (B) data before PSM and Blood glucose (C), and fluctuation(D) data after PSM. Blood glucose mainly contains FBG, PBG, and BG and blood fluctuation mainly contains SDBG, PBGE, and LAGE. Blood glucose data collected from blood glucose monitoring from fasting blood glucose (FBG, 6:00 am), post-breakfast blood glucose (PBG, 9:00 am), post-lunch blood glucose (PBG,13:00 pm), and post-dinner blood glucose (PBG,19:00 pm). FBG data were collected 3 days before discharge and PBG data were collected 24 h before discharge and fluctuation data were calculated mainly based on 24 h before discharge. BG, blood glucose; LAGE, largest amplitude of glycemic excursions; PBGE, postprandial glucose excursion; SDBG, standard deviation of blood glucose. Data were described as mean ± SE, p < .05 was considered as a significant difference. ** p < .01 and * p < .05 TA B L E 2 Fasting blood glucose stratify according to blood glucose standards before propensity score matching Blood glucose data collected from blood glucose monitoring from fasting blood glucose (FBG, 6:00 am), FBG data were collected 3 days before discharge. Frequency of FBG<7.8 mmol/L and FBG 7.8-10 mmol/L in 3 days before discharge. Data were described as mean ± SE, p < .05 was considered as a significant difference.
BG before and after the pharmacist consultations, but also evaluated these factors between the pharmacist intervention and nonintervention patients. We found that the blood glucose indices (FBG, PBG, and mean BG) were significantly reduced after the pharmacist intervention. In order to further verify our findings, we used the PSM method to control for multiple patient factors. The PSM result showed that the blood glucose indices improved with the pharmacist interventions.
Glycemic disorders, including diabetes, impaired glucose tolerance and stress hyperglycemia are commonly seen in patients with AMI. 19 Glycemic disorder is considered a major predictor of poor clinical outcomes in AMI patients, which is not only limited to constant hyperglycemia, but also involves glucose fluctuations. 20,21 Most previous studies mainly concentrated on HbA1c and other cardiovascular risk factors after AMI. 11,22 There is a lack of as-   28 and possibly, poor knowledge of some specific drugs. 29

| Strengths and limitations
The strengths of this study relate to the pharmacist intervention in diabetic patients with AMI. Professional endocrine clinical pharmacists not only designed the medical schedule, but also monitored the implementation of the plan, followed up with patients, and continuously optimized the scheme. The PSM analysis was also strength. Given that the baseline characteristics (including HbA1c%, FBG, PBG, and longer length of hospital stay) were not comparable between the pharmacist intervention and nonintervention groups, the PSM method was used to reduce the confounding factors, resulting in 19 patients in each group. Several limitations in this preliminary study need to be considered. First, this was a single-center study. Second, the sample size was insufficient. Therefore, further studies with larger sample sizes and multiple centers are needed to strengthen the conclusion that the pharmacist intervention model improves blood glucose management. Furthermore, this model was applied mainly within the department of cardiology and cardiac care unit in our institution.
The extrapolation of this model can be further standardized by establishing an application (app) or WeChat program. Third, although we found that the pharmacist intervention was beneficial in controlling blood glucose management and mitigating glucose fluctuation, we did not include clinical events in this study. Finally, we calculated the indices of the blood glucose fluctuation using finger-prick blood glucose data. A previous study 30 suggested that the traditional finger-prick blood glucose level monitoring provided a relatively reasonable approximation of the mean blood glucose concentration in most patients. However, this may underestimate the prevalence of potential hyperglycemia and hypoglycemia. Therefore, in future studies, the use of continuous blood glucose monitoring is recommended and may further verify our conclusions.

| CON CLUS ION
In this study, we found that the pharmacist intervention improved the blood glucose levels (mean FBG, mean PBG, and mean BG) as well as glucose fluctuations (SDBG, PBGE, and LAGE). Furthermore, clinical pharmacists may reduce the potential risk for hypoglycemia and make the blood glucose control more achievable for diabetic patients with AMI. The treatment model coordinated by clinical pharmacists and the clinicians may be recommended for blood glucose control, especially for patients with cardiovascular complications. This collaborative treatment model, involving clinical pharmacists and clinicians, should be the trend for future developments. Experiences based on this preliminary study were limited to a specific group of patients, but this study may serve as an example of a promising approach for blood glucose management.

ACK N OWLED G EM ENTS
Gu and Li are the guarantors of the entire manuscript. Shi and Shen contributed to the study conception and design, critical revision of the manuscript for important intellectual content, and final approval of the version to be published. Yue contributed to the data acquisition, analysis, and interpretation. Ma and Lin contributed to supervised the investigation and revised the manuscript.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflicts of interest for this article.

E TH I C S A PPROVA L S TATE M E NT
This study protocol was approved by ethics committees of Renji Hospital, School of Medicine, Shanghai Jiaotong University (KY2019-076).

PATI E NT CO N S E NT S TATE M E NT
Informed consent was obtained from each patient or their family members.

PR I N CI PA L I N V E S TI G ATO R S TATE M E NT
The authors confirm that the principal investigator for this paper is Fang-Hong Shi and Zhi-Chun Gu and that both of them had direct clinical responsibility for patients.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data supporting the conclusion of this article will be made available by the authors to related qualified researcher.