Predicted resting metabolic rate and prognosis in patients with ischemic stroke

Abstract Purpose Resting metabolic rate (RMR) could represent metabolic health status. This study aims to examine the association of the predicted RMR with 1‐year poor functional outcome and all‐cause mortality in patients with ischemic stroke as a proxy of metabolic profile. Methods A total of 15,166 patients with ischemic stroke or transient ischemic attack (TIA) from the Third China National Stroke Registry (CNSR‐III) were enrolled in this study. The Harris–Benedict equation based on sex, age, weight, and height was used to predict RMR. The primary endpoints were poor functional outcome defined as ≥3 modified Rankin Scale (mRS) score and all‐cause mortality within 1 year. The association between predicted RMR and prognosis was assessed by multivariable regression analysis. Besides that, subgroup analysis of age, sex, and body mass index (BMI) with predicted RMR was also performed. Results 12.85% (1657) individuals had poor functional outcome and 2.87% (380) died of whatever causes within 1 year. An inverse association was found between predicted RMR with poor functional outcome and all‐cause mortality. Compared to the lowest quartile, the highest quartile was significantly associated with lower risk of poor functional outcome (adjusted odds ratio [OR], 0.43 [95% confidence interval (CI) 0.33–0.56]) and all‐cause mortality (adjusted hazard ratio [HR], 0.44 [95% CI 0.28–0.71]). No significant interaction was between predicted RMR and specified subgroup. Conclusions Predicted RMR by the Harris–Benedict equation seems to be an independent protective predictor of poor functional outcome and all‐cause mortality after ischemic stroke as a metabolic proxy.


INTRODUCTION
Ischemic stroke is the leading cause of physical disability and causes enormous economic burdens . It is imperative to recognize the related markers and factors associated with poststroke results in arrange to direct clinical stroke administration way better.
Emerging evidence suggested that metabolic related factors, such as obesity, insulin resistance, type 2 diabetes mellitus, have a great implication for stroke incidence as well as its prognosis (Lau et al., 2019;Rodríguez-Castro et al., 2019;Yang et al., 2021).
Resting Metabolic Rate (RMR), defined as the daily energy required to ensure life-sustained functions in resting status, takes up 60-70% of the total daily energy demands (Zampino et al., 2020). Age, height, weight, sex, genetic variation, and physiology all are responsible for RMR. Previous studies indicated that RMR may represent the metabolic health status as a marker of whole body metabolism level.
For example, a recent report from the European Prospective Investigation into Cancer and Nutrition (EPIC) showed that the higher basal metabolic rate estimated by WHO/FAO/UNU equation was associated with a greater risk of specific cancers (Kliemann et al., 2020). Another supportive example was that the higher metabolic rate assessed using indirect calorimetry an older adult had, the greater burden of multimorbidity he/she might have (Fabbri et al., 2015). That indicated that the RMR may be an indicator for prognosis after ischemic stroke as a proxy of metabolic profiles same as for cancer or natural death (Kliemann et al., 2020;Ruggiero et al., 2008). However, the effects of RMR on poststroke prognosis have not been investigated in depth. It is worth investigating the relationship between them in terms of better prognosis management of ischemic stroke.
This study aims to assess the association between predicted RMR as a marker of metabolic profile and the prognosis of ischemic stroke based on the Third China National Stroke Registry (CNSR-III) database . Owing the initial design of the CNSR-III database did not include the RMR variable measured by indirect calorimetry, we used the predicted RMR by the Harris-Benedict equation as an alternative.

Study design
This study was based on baseline and follow-up data from the CNSR-III study. The rationale, design, and baseline participant characteristics of the CNSR-III had been described previously  (3) cancer at recruitment; (4) discharge diagnosis of TIA.

Standard protocol approvals, registrations, and patient consents
The study was approved by the ethics committees of Beijing Tiantan Hospital (IRB approval number: KY2015-001-01) and all participating centers. Written informed consents were signed by all participants or their legal proxies before enrolling in the study.

Assessment of the predicted resting metabolic rate
Harris-Benedict equation derived early in 1919 is one of the most frequently used to predict RMR in clinical application (Harris & Benedict, 1918). This method calculates RMR using sex-specific equations and is also based on the participant's age, weight, and height. Accumulating evidence indicated that predicted RMR by the Harris-Benedict equation was associated well with that by indirect calorimetry in nonobesity, healthy participants and even more so in those with obesity (Bendavid et al., 2021). We chose the Harris-Benedict equation to assess RMR. Meanwhile, RMR was also calculated by using the other  Table S1) (Energy & Protein Requirements, 1985;Henry, 2005;Kliemann et al., 2020;Mifflin et al., 1990).

Outcome and follow-up
Follow-up time started from the day of enrollment in the registry project. Patients were followed up for 1 year after ischemic stroke by trained research coordinators over the telephone. Information on functional status and all-cause death was collected. Each case death was confirmed according to a death certificate from the attended hospital or the local citizen registry. The primary study endpoints included poor functional outcome, defined as functional dependency based on mRS score of 3 to 5, and all-cause mortality resulting from the index event or other causes in 1-year follow-up.

Statistical analysis
The predicted RMR at baseline was calculated by the Harris-Benedict confounders due to certain potential factors (Supplemental Table S3).
The restricted cubic spline model indicated a similar associated pattern to the predicted RMR as a categorical variable both in the relationship with poor functional status and all-cause mortality (Figure 2).
No significant heterogeneity was observed across specified subgroups of age, sex, and BMI on the primary endpoints (all p for interaction >.05; Supplemental Table S4). The sensitivity analysis performed in 12,914 participants of mRS of 0 to 3 did not change the above reverse associations (Supplemental Table S5).

DISCUSSION
This study provided that after controlling for related confounders, pre- that predicted RMR seemed to be an independent protective predictor for prognosis of ischemic stroke as a proxy of metabolic profile.
Given that the RMR measured in current study was only based on the age, weight, and height variable by sex, so in theory, the male subjects with bigger body size (greater weight and higher) and younger age were more likely to obtain good outcome after ischemic stroke.
Although it seemed to appear that outcome status was decided on the pure sum of the variables in equation, we should note that each of the variables has different weighted factors which indicted that different RMR would be obtained by changes in same unit of every variable.
Owning to measurements on RMR obtained by Harris-Benedict equation (closely associated with indirect calorimetry in nonobesity, healthy participants and most subjects in this study satisfied these conditions as presented in Table 1) instead of indirect calorimetry, with consideration of its significance in clinical practice, further testified studies on stroke and RMR with an indirect calorimetry method were expected.
Using RMR as a potential biomarker of body metabolic health has been supported by several previous studies. However, the conclusion drawn on the dangerous indicator of higher RMR for diseases or death appeared to predominate over its protective side (Drabsch et al., 2018;Jumpertz et al., 2011;Kang et al., 2021;Kliemann et al., 2020;Ruggiero et al., 2008;Schrack et al., 2014). An over 40-year follow-up of 1227 healthy participants conducted in the Baltimore Longitudinal Study of Aging (BLSA), mostly men, indicating those with high basal energy rate was associated with shorter longevity. (Ruggiero et al., 2008) That was supported by an over 2-year calorie restriction trial on 53 nonobese adults showing that decreased energy expenditure improved the rate of living. (Redman et al., 2018) Higher incidences of urolithiasis recurrence and diabetes were observed in the individuals with higher RMR (Kang et al., 2021). Nonetheless, diabetes and metabolic syndrome were also found more likely to occur in the population with low RMR (Buscemi et al., 2007;Georgopoulos et al., 2009;Maciak et al., 2020;Olive et al., 2008). Our results seemed to disagree with the tendency of the dangerous implication of RMR for healthy state. The target population among studies, like whether they were healthy, sex-specific dominant, obese, with ischemic stroke or not, had different demographic characteristics. In addition, difference in study design and RMR evaluating methods among studies could partly account for the discrepancy of conclusions. Our interests in terms of secondary prevention instead of primary prevention could cause the appearance of protective effect that may be similar to the relationship between obesity and ischemic stroke (Vemmos et al., 2011). The controversial relationship between RMR and oxidative stress that is considered to damage health might be one of the inconsistent causes (Frisard & Ravussin, 2006). Therefore, to confirm our findings, further studies evaluating populations affected by stroke or other diseases are warranted.
A consistent tendency between the level of predicted RMR and BMI (p < .0001) in our study may result from the common crucial variable of weight (Table 1). In the subgroup analysis of overweight/obesity, our results found predicted RMR could act as an independent protective factor for the poor functional outcome (p for interaction, .24) and allcause mortality (p for interaction, .84) (Supplemental Table S3). This indicated that predicted RMR could detect extra information beyond BMI, while BMI is positively associated with improved outcome after ischemic stroke (Vemmos et al., 2011).
Although scarce studies directly examined the protective predictive value of RMR on unfavorable ischemic stroke prognosis, several mechanisms could account for the inverse association. The positive relationship between RMR and cardiorespiratory fitness has been found, and the latter is treated as a target to reduce cardiovascular diseases and mortality. Higher RMR could represent better cardiorespiratory fitness and promotes functional recovery (Ebaditabar et al., 2021;Shook et al., 2014). On the other hand, higher RMR represents greater composition of skeletal muscle, the loss of which significantly influences the physical activity Oh et al., 2019;Soysal et al., 2019;Visser et al., 2005). Additionally, a series of spontaneous recovery processes after ischemic stroke, such as the activity of neuroglial cells and the changes of repair-related molecular, need support of the strong reverse capacity that may be along with great RMR (Bélanger et al., 2011;Cramer, 2008;Zampino et al., 2020).
The study first applied a large-scale cohort with ischemic stroke to examine the association between energy metabolism and ischemic stroke prognosis. However, some limitations should be considered.
First, the RMR was predicted by the equation instead of a gold standard with indirect calorimetry owing to the expensive prices and complex operation. A decrease in accuracy of the Harris-Benedict equation may occur when used in individuals with low or high BMI and older age (Jésus et al., 2015). Given that, we excluded the patient whose age or BMI was beyond the range from 30 to 80 and from 16 to 40, respectively. Second, the predicted RMR was measured once at baseline regardless of its potential change over time. Third, the information on the self-reported thyroid disease lacked in our analysis. Concerning F I G U R E 2 Restricted cubic spline for poor functional outcome (A) and all-cause mortality (B) according to levels of predicted RMR. Solid dark line is a multivariable adjusted OR/HR, with a gray area indicating 95% CI derived from restricted cubic spline regression. A dashed dark line represents no association.
that the prevalence of overt hyperthyroidism and hypothyroidism in a nationwide survey with an enrollment of 78, 470 Chinese adults was 0.78% (95% CI, 0.69% to 0.87%) and 1.02% (95% CI, 0.88% to 1.18%) respectively, our results might not be influenced substantially . Fourth, there were a lot of potential confounders like nutritional status, body composition and etc. which we did not assess due to lack of necessary related data, that could interfere the outcomes. Finally, relatively short follow-up period and small cumulative number of deaths (380, 2.87%) led to no death in female patient with highest predicted RMR, which could cause a lower statistical power in assessing the association between predicted RMR and outcome.
In summary, higher predicted RMR could be an independent protective indicator for the risk of poor functional outcome and all-cause mortality of ischemic stroke as a proxy of metabolic profile. Nevertheless, it is not yet known whether the RMR is merely a proxy or plays a causal role in the relationship with poststroke outcome. Figuring it out will facilitate to improve outcome and decrease death and disability by changing the RMR. We here proposed the following potential points regarding the further study on RMR: use of the more rigorous methods to verify the association like golden standard, study of the molecular and cellular mechanisms that influence RMR so as to seek to transform it into pharmaceutical and behavioral intervention.

AUTHOR CONTRIBUTIONS
XYL and ACC drafted the manuscript and interpreted the data. YSP and MXW contributed to revised the manuscript and statistical problems. XM revised the manuscript. YJW interpreted data and revised the manuscript. All authors read and approved the final manuscript to be published.

ACKNOWLEDGMENTS
We thank all the staff and participants of the CNSR-III (Third China National Stroke Registry) studies for their contribution.

CONFLICT OF INTEREST STATEMENT
The author declares that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

DATA AVAILABILITY STATEMENT
All data are available to researchers on request for purposes of reproducing the results or replicating the procedure by directly contacting the corresponding author.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/brb3.2911