Loading [MathJax]/jax/output/SVG/jax.js
Review

The underlying mechanisms for severe COVID-19 progression in people with diabetes mellitus: a critical review

  • Diabetes mellitus (DM) has a high incidence of comorbidities among patients with severe coronavirus disease 2019 (COVID-19). The elevated prevalence of DM in the world population makes it a significant risk factor because diabetic individuals appear to be prone to clinical complications and have increased mortality rates. Here, we review the possible underlying mechanisms involved in DM that led to worse outcomes in COVID-19. The impacts of hyperglycemia side effects, secondary comorbidities, weakened innate and adaptive immunity, chronic inflammation, and poor nutritional status, commonly present in DM, are discussed. The role of the SARS-CoV-2 receptor and its polymorphic variations on higher binding affinity to facilitate viral uptake in people with DM were also considered. Clinical differences between individuals with type 1 DM and type 2 DM affected by COVID-19 and the potential diabetogenic effect of SARS-CoV-2 infection were addressed.

    Citation: María D Figueroa-Pizano, Alma C Campa-Mada, Elizabeth Carvajal-Millan, Karla G Martinez-Robinson, Agustin Rascon Chu. The underlying mechanisms for severe COVID-19 progression in people with diabetes mellitus: a critical review[J]. AIMS Public Health, 2021, 8(4): 720-742. doi: 10.3934/publichealth.2021057

    Related Papers:

    [1] Suqing Lin, Zhengyi Lu . Permanence for two-species Lotka-Volterra systems with delays. Mathematical Biosciences and Engineering, 2006, 3(1): 137-144. doi: 10.3934/mbe.2006.3.137
    [2] Yuanshi Wang, Donald L. DeAngelis . A mutualism-parasitism system modeling host and parasite with mutualism at low density. Mathematical Biosciences and Engineering, 2012, 9(2): 431-444. doi: 10.3934/mbe.2012.9.431
    [3] Mengqing Zhang, Qimin Zhang, Jing Tian, Xining Li . The asymptotic stability of numerical analysis for stochastic age-dependent cooperative Lotka-Volterra system. Mathematical Biosciences and Engineering, 2021, 18(2): 1425-1449. doi: 10.3934/mbe.2021074
    [4] S. Nakaoka, Y. Saito, Y. Takeuchi . Stability, delay, and chaotic behavior in a Lotka-Volterra predator-prey system. Mathematical Biosciences and Engineering, 2006, 3(1): 173-187. doi: 10.3934/mbe.2006.3.173
    [5] Sheng Wang, Lijuan Dong, Zeyan Yue . Optimal harvesting strategy for stochastic hybrid delay Lotka-Volterra systems with Lévy noise in a polluted environment. Mathematical Biosciences and Engineering, 2023, 20(4): 6084-6109. doi: 10.3934/mbe.2023263
    [6] Xiaomeng Ma, Zhanbing Bai, Sujing Sun . Stability and bifurcation control for a fractional-order chemostat model with time delays and incommensurate orders. Mathematical Biosciences and Engineering, 2023, 20(1): 437-455. doi: 10.3934/mbe.2023020
    [7] Monika Joanna Piotrowska, Urszula Foryś, Marek Bodnar, Jan Poleszczuk . A simple model of carcinogenic mutations with time delay and diffusion. Mathematical Biosciences and Engineering, 2013, 10(3): 861-872. doi: 10.3934/mbe.2013.10.861
    [8] Xixia Ma, Rongsong Liu, Liming Cai . Stability of traveling wave solutions for a nonlocal Lotka-Volterra model. Mathematical Biosciences and Engineering, 2024, 21(1): 444-473. doi: 10.3934/mbe.2024020
    [9] Junjing Xiong, Xiong Li, Hao Wang . The survival analysis of a stochastic Lotka-Volterra competition model with a coexistence equilibrium. Mathematical Biosciences and Engineering, 2019, 16(4): 2717-2737. doi: 10.3934/mbe.2019135
    [10] Jinyu Wei, Bin Liu . Global dynamics of a Lotka-Volterra competition-diffusion-advection system for small diffusion rates in heterogenous environment. Mathematical Biosciences and Engineering, 2021, 18(1): 564-582. doi: 10.3934/mbe.2021031
  • Diabetes mellitus (DM) has a high incidence of comorbidities among patients with severe coronavirus disease 2019 (COVID-19). The elevated prevalence of DM in the world population makes it a significant risk factor because diabetic individuals appear to be prone to clinical complications and have increased mortality rates. Here, we review the possible underlying mechanisms involved in DM that led to worse outcomes in COVID-19. The impacts of hyperglycemia side effects, secondary comorbidities, weakened innate and adaptive immunity, chronic inflammation, and poor nutritional status, commonly present in DM, are discussed. The role of the SARS-CoV-2 receptor and its polymorphic variations on higher binding affinity to facilitate viral uptake in people with DM were also considered. Clinical differences between individuals with type 1 DM and type 2 DM affected by COVID-19 and the potential diabetogenic effect of SARS-CoV-2 infection were addressed.


    Abbreviations

    DM:

    Diabetes mellitus; 

    COVID-19:

    Coronavirus disease 2019; 

    SARS-CoV:

    Severe Acute Respiratory Syndrome Coronavirus; 

    SARS-CoV-2:

    Severe Acute Respiratory Syndrome Coronavirus 2; 

    CIAD:

    Research Center for Food and Development; 

    WHO:

    World Health Organization; 

    CSSE:

    Center for Systems Science and Engineering; 

    MERS:

    Middle East respiratory syndrome; 

    ICU:

    Intensive care unit; 

    ACE1:

    Angiotensin-converting enzyme 1; 

    ACE2:

    Angiotensin-converting enzyme 2; 

    ACEi:

    ACE inhibitors; 

    RAAS:

    Renin-angiotensin-aldosterone system; 

    ARBs:

    Angiotensin-receptor blockers; 

    HbAc1:

    glycosylated hemoglobin; 

    ROS:

    Reactive oxygen species; 

    CD4+:

    Cluster of differentiation 4+; 

    CD8+:

    Cluster of differentiation 8+; 

    TNF-α:

    Tumor necrosis factor-α; 

    IL:

    Interleukin; 

    CRP:

    C-reactive protein; 

    IFN-ϒ:

    Interferon-ϒ; 

    CT:

    Chest computed tomography; 

    LHD:

    Lactate dehydrogenase levels; 

    RBD:

    Receptor-binding domain; 

    RNA:

    Ribonucleic acid; 

    DNA:

    Desoxyribonucleic acid; 

    SNPs:

    Single nucleotide polymorphisms; 

    CHD:

    Coronary heart disease; 

    BMI:

    Body mass index; 

    DKA:

    Diabetic ketoacidosis; 

    TMPRSS2:

    transmembrane serine protease 2

    With the popularization of cloud computing [1], cybersecurity has become a major concern [2,3]. According to the model of service, cloud computing can be divided into public, private and hybrid cloud. The private cloud is deployed for certain organizations with higher requirements in terms of performance and security. Virtual machine (VM), as the main form providing service to users in the cloud environment, has increasingly become the target of cyber-attacks, leveraging malwares to comprise the VM system and leading to significant damages to the virtual and even physical platforms.

    There are many kinds of malwares such as virus, worm, Trojan and backdoor. As one of the most dangerous malwares, kernel rootkit [4], also called Driver Trojan, obtains the system administrator privilege while hiding its existence in the OS kernel. Kernel rootkits modify the critical kernel data structures, making it much more difficult for detection than any other kinds of malwares. Traditional anti-virus solutions are based on the signatures of the already known malwares whilst installed in the same system with the malware. They can only detect known malwares and can be easily spotted and evaded by the smart malwares.

    Automatic detection of known and unknown kernel rootkits on VMs is becoming an urgent issue. As the kernel rootkit is hard to detect in the user mode, we observe its tracks from the memory side. As a kind of digital investigation for detecting cybercrimes [5], memory forensic can effectively extract malicious features from the memory dumps of the monitored computer. It is especially suitable for kernel rootkits detection. In virtualization environment, the memory dump acquisition of the VMs can be conducted out of the monitored VM at the higher hypervisor level in a trusted way, so the memory acquisition operation cannot be discovered or subverted by malwares. Machine learning technology has proved effective at automatically detecting known and unknown malware in many researches [6,7,8]. We combine the machine learning technology with memory forensic analysis to form a novel kernel rootkit detection method TKRD for the VMs in private cloud.

    The main contribution of this paper can be highlighted as follows:

    ● A novel TKRD method to automatically detect kernel rootkits in VMs of private cloud was proposed;

    ● The memory dumps of the VM were acquired from the controller node of the OpenStack platform in a trusted way;

    ● Memory forensic analysis along with the Volatility framework was leveraged to extract malicious features;

    ● Bio-inspired machine learning classifiers were trained for automatic malware detection;

    ● The experiment results showed that our methodology can effectively detect the known and unknown kernel rootkits for VMs.

    The rest of the paper is organized as follows. Related works are introduced in Section 2. Section 3 presents the memory acquisition and machine learning methods used in our research. The experimental settings and results are discussed in Section 4, followed by some concluding remarks summarized in Section 5.

    Rootkits are malwares allowing permanent or consistent, undetectable presence in a computer system [9]. Rootkits can hide specific system resources to achieve the goal of hiding the intrusion into the compromised computer. They are categorized into five classes: user mode rootkits, kernel mode rootkits, bootkits, hypervisor level rootkits and firmware rootkits. User mode rootkits run on Ring 3 of the x86 processor with the least privilege. Kernel mode rootkits execute in privileged mode on Ring 0, making it very hard to detect. Bootkits are a variant of kernel mode rootkits, which can infect boot code such as Master Boot Record (MBR), Volume Boot Record (VBR), or boot sector. Hypervisor level rootkits exploit hardware virtualization features running in Ring -1 which hosts the VM operating system (OS). Firmware rootkits exist in the firmware of hardware such BIOS. As the integrity of firmware is not checked so frequently, it is an ideal place for firmware rootkits to hide. In this paper we focus on the kernel mode rootkits.

    Rootkits leverage many techniques to subvert the OS. The flow of the program can be modified by installing hooks. For example, by manipulating function pointers in System Service Descriptor Table (SSDT), the transform from user mode API to system Native API is compromised. More advanced rootkits such as FU and Shadow Walker can launch Direct Kernel Object Manipulation (DKOM) attacks, which directly modify the core data structure of OS kernel in memory. Malicious library injection and code injection are also common means for rootkits to subvert the system.

    Memory forensic belongs to the rank of digital forensics which collect and present digital evidence for cybercrime investigation. There are many researches on leveraging memory forensic to uncover the malicious behaviors in the memory.

    Case et al. [10] presented a new kernel rootkit detection technique suitable for Mac OS X system. They used the most popular memory forensic framework Volatility to analyze the features of malwares. For each new rootkit detection technique, a Volatility plugin was developed. Evaluation of the plugins were done to illustrate their effectiveness.

    For detecting malwares in Android, Yang et al. [11] proposed a general tool named "AMExtractor" for volatile memory acquisition for Android devices. Using memory forensic tools, the memory dumps acquired were further analyzed. Rootkits were successfully detected by memory forensic framework Volatility.

    For malware detection in virtualization environment, Kumara et al. [12] leveraged memory forensic tool such as Volatility, Rekall to analyze the memory state of the VMs, which can address the semantic gap problem existing in Virtual Machine Introspection (VMI). LibVMI can also address the semantic gap problem, and eliminate memory dump acquire time, achieving a high efficiency of memory dump acquisition and analysis.

    Using memory forensic technology, Hua et al. [13] designed and implemented a VMM(virtual machine monitor)-based hidden process detection system to investigate rootkits by identifying the lack of the critical process and the target hidden process from the aspect of memory forensics. They addressed the limitation that the tradition host-based detection tools could be deceived or subverted by malwares.

    Tien et al. [14] introduced a memory data monitoring method against the running malware outside the VM, various features were observed from the memory. At last, they suggested combining memory forensic with heuristic method to promote the detection rate for malwares. From the above researches, memory forensic is an effective way to detect complex malwares, especially rootkits.

    Machine learning technology has proved effective in automatically detecting known and unknown malwares in many researches. Cohen et al. [15] proposed a method for trusted detection of ransomware in private cloud. Memory forensic framework Volatility was used to extract meta-features from the volatile memory dumps of the VM. These meta-features were used to train machine learning classifiers. The experiment results showed that the method can effectively detect unknown ransomware.

    In order to automatically detect malware in cloud VMs, NirNissim et al. [16] conducted a research on memory forensic for cloud VM. System-call sequences were extracted from the memory dumps as features for malware detection. Various machine learning algorithms were leveraged to detect malwares in the cloud. They focused on the ransomware and RAT malware in IIS and Email server. The experiment results showed that the methodology can detect unknown ransomware and RAT in an average 97.9% true positive rate (TPR) and 0% false positive rate (FPR).

    Kumara et al. [17] proposed an automated multi-level detection system for Rootkits and other malwares for VMs at hypervisor level. The detection system was composed of Online Malware Detector (OMD) and Offline Malware Classifier (OFMC). The OMD detected known malwares based on the local malware signature database. The OFMC classified the unknown malwares by adopting machine learning technique. The meta-data for classifier training was extracted by virtual machine introspection tools from the running VMs at the hypervisor level. The experiments achieved 100% Accuracy and zero FPR. Himanshu et al. [18] proposed an automated and real-time monitoring and forecasting system for rootkits and other malwares for the Windows virtualization architecture. It firstly extracted kernel data structures with VMI tool LibVMI. Then a prediction tool was developed using machine learning techniques. However, the authors did not tell which data structure to extract and how to train the machine learning classifier.

    Mosli et al. [19] proposed an automated malware detection method using artifacts in forensic memory images. The features extracted from the memory images were registry activity, imported libraries, and API function calls. Different machine learning techniques were implemented to compare performances of malware detection. The highest Accuracy was reached by a SVM model with the feature of registry activity. Ajay Kumara et al. [20] proposed an advanced VMM-based machine learning techniques at the hypervisor. Machine learning techniques were used to analyze the executables that were mined and extracted using MFA (memory forensic analysis)-based techniques and ascertained the malicious executables. The evaluation results achieved 99.55% Accuracy and 0.4% FPR on the 10-fold cross-validation to detect unknown malware on the generated dataset. Bai et al. [21] improved malware detection using multi-view ensemble learning. Because malwares were hard to disguise itself in every feature view, so they designed two schemes to incorporate three single-view features to fully exploit complementary information to discover malwares which achieved a false alarm rate of 0%.

    In our work, we combine machine learning technology with memory forensic to detect kernel rootkits in VMs. Firstly, VM memory dumps are acquired in a trusted way at the hypervisor level. Then memory forensic analysis framework Volatility is used to extract malicious features from the memory dumps. Finally, based on these features, machine learning classifiers are trained. We will describe the dataset acquisition and machine learning classifier training in detail in the next section.

    In this section, the methods used in our research are presented. In subsection 3.1, we introduce the way to obtain the volatile memory dumps from the monitored VM. In subsection 3.2, the methods of feature extraction and representation are introduced. In subsection 3.3, we describe machine learning algorithms used to train classifiers. The flowchart of the methods is shown in Figure 1.

    Figure 1.  The flow chart of the malware detection.

    OpenStack [22] is a popular cloud operating system, which manages the whole cloud computing environment. Users utilize OpenStack to deploy their public, private and hybrid cloud environment. In our research, we leverage OpenStack to deploy a private cloud experiment environment. OpenStack manages various resources throughout datacenter and provides dashboard or API to users to simplify the routine operations. The topological structure of the virtualization management infrastructure is shown in Figure 2.

    Figure 2.  The topological structure of OpenStack platform.

    The network node provides network service for the whole platform. The controller node is responsible for managing other nodes, providing identity service, image service and dashboard service. The storage node includes block storage service and object storage service. The virtual machine instances are created on the compute nodes, which provide computing service. By default, the kernel-based VM (KVM) hypervisor is supported by OpenStack. So we choose KVM as our hypervisor. Windows 7 Ultimate (x86) has been installed on the VM.

    We use the dashboard on the controller node to create VM instances on the compute node. The dashboard supplies snapshot acquisition function. During the VM running process, we get its snapshots including volatile memory dumps and other information. We only extract the volatile memory dump documents for data analysis. Because the memory dumps are obtained out of the monitored VM, the malware running in the VM cannot evade or interfere with the memory obtaining process. Otherwise, the memory acquisition mechanism may be destroyed or evaded by the smart malwares, if it is located in the same VM with the malware. So it is a trusted way to gain the memory dumps of the VM.

    To collect training and testing dataset, ~100 kernel rootkit samples belonging to the family of NTRootkit, FURootkit etc. and their variants have been downloaded from the malware repository. The rootkit samples run on the VM one at a time. As the malware is dynamic, the state of the memory will be constantly changing. During the running time of each sample, we capture its memory states at appropriate intervals, for example every 10 minutes. Because there are long running samples and short running samples, the interval for memory capture are varied. For the short running samples the intervals are shorter to avoid the issue that they bypass the detection. The intervals of memory capture can be short or long depending on the requirement of the system. To this end, this sampling interval can be adaptively set as a tradeoff of efficiency and efficacy. For each sample, 100 memory dumps have been captured. The total number of captured memory dumps is 10000. Malicious features of each dump have been extracted, forming a record in the dataset. To get the benign memory samples, we captured memory state of the fresh installed Windows OS and some commonly used applications such as office, PDF, Winzip, Google browser etc.. During the running process of each application, the memory dumps have been obtained in the same way as the malicious. The total number of the benign memory dumps is 2300. An automatic tool for memory dump acquisition and feature extraction has been developed based on Python script.

    After obtaining the memory dumps, features need to be extracted from them for abnormal behavior detection. Memory forensic framework Volatility [23] is used to extract features for memory forensic analysis.

    As a widely used memory forensic platform, Volatility is a completely open collection of tools, implemented in Python under GNU Public License. Memory forensics for Windows, Linux and MAC operating systems are supported by Volatility. In our research, Volatility 2.6 is leveraged to extract digital artifacts from volatile memory samples of Windows 7. Various plugins are supplied by Volatility to extract information from memory dumps. In our research, the following plugins are utilized to extract features, shown in Table 1.

    Table 1.  Volatility plugins used for rootkit detection.
    Volatility plugin Description
    modules Print loaded modules
    modscan Pool scanner for kernel modules
    threads Investigate _ETHREAD and _KTHREADs
    OrphanThread Investigate orphan threads
    driverscan Pool scanner for driver objects
    driverirp Pool scanner for driver objects
    devicetree Show device tree
    ssdt Display SSDT entries
    callbacks Print system-wide notification routines
    timers Print kernel timers and associated module DPCs

     | Show Table
    DownLoad: CSV

    According to the way rootkit attacks, the features to be extracted are as follows [24]: hidden kernel modules, orphan threads, driver objects, device tree, the SSDT function, call backs and timers.

    Hidden kernel modules: The rootkit loads into the kernel as a kernel module. In order to hide its existence, rootkit hides its module. A metadata structure KLDR_DATA_TABLE_ENTRY is generated when a kernel module is loaded into the kernel. All of the metadata structures form a doubly linked list. The rootkit can hide its presence by unlinking the corresponding entry from the list, as shown in Figure 3.

    Figure 3.  The doubly linked list of module entry.

    In Volatility, the modules plugin walks through the double linked list to enumerate the kernel modules. If the metadata structure is unlinked, the module becomes stealthy for the modules plugin. Although the metadata entry is unlinked from the list, it is still in the memory, which is tagged with MmLd. The modscan plugin uses pool-scanning approach to find the metadata entry by the tag of Mmld. Through comparing the scanning results of these two plugins, the unlinked stealthy modules can be found.

    Orphan threads: Orphan threads do not belong to any modules, which is an abnormal phenomenon in memory. For the purpose of hiding their presence, some rootkits unload the module when the executing thread is created. It can evade the detection of suspicious modules, but will create Orphan thread which is an obvious artifact for rootkit detection. We use threads and OrphThread plugins to find out orphan threads.

    Driver objects: When a kernel module is loaded, a corresponding structure _DRIVER_OBJECT is initialized. The driver object contains information about the kernel module. Even when the tag MmLd of the metadata structure KLDR_DATA_TABLE_ENTRY is corrupted, the driverscan plugin can still detect the module. So through the comparison between the results of modules and driverscan plugins, we can discover the hidden module whose metadata structure has been destroyed.

    Every driver object contains a major function table that handles different requests from the OS user mode. Rootkit can hook these functions. To discover the hooks, driverirp plugin is needed to display the 28 values in the MajorFunction array. Of course, further analysis is needed to decide the existence of the hooks.

    Devicetree: In Windows OS, one device object is permitted to attach to the stack of another device. So they can handle the same I/O request, achieving transparent system archiving and encryption. This also provides rootkits with opportunities to intercept the legitimate I/O Request Packets (IRPs). The devicetree plugin can be used to audit the device tree to inspect whether there are malware drivers attaching to the legitimate device.

    SSDT: SSDT containing pointers to kernel mode functions is a big target for rootkits. Many rootkits get to execute malicious modules by modifying or hooking the SSDT. So it is necessary to audit the SSDT. There are several ways to attack the SSDT functions. Some rootkits overwrite the pointers in SSDT. Normally, SSDT only points to two tables: native API table and shadow SSDT table, corresponding to the module of ntoskrnl.exe and win32k.sys respectively. The ssdt plugin can detect the modified pointers pointing to other malicious modules. More cunning rootkits use inline hook technique, instead of pointing out of the NT modules or win32k.sys. In this case, further static analysis is needed.

    Callbacks: The callback plugin can show all the callbacks installed on the system, including information about the callback type, the start address of the callback function, the corresponding module of the callback function and some other information. If a rootkit has hidden its module, the corresponding module information will show unknown, which is an obvious abnormal artifact. Take one step back, some rootkits don't hide their modules, but use a hard-coded name which also indicates a compromise.

    Timers: Timer is another feature indicating a rootkit. Each timer is associated with a module. The timers plugin can list all timer objects created in the memory, including the module that the timer belongs to. So if the module shows unknown, the timer is suspicious.

    The above is all the features used to detect rootkits in our research. Different rootkits may have different features. In Table 2, we list the features and the plugins extracting them. We develop a Python-based tool to automatically extract these features from the memory dumps.

    Table 2.  The features extracted and plugins used.
    Feature Description Data type Plugin
    Hidden modules Whether hidden modules exist Boolean modulesmodscan
    Orphan thread Whether orphan threads exit Boolean threadOrphanThreads
    Driver object Whether abnormal driver object exits Boolean driverscan modules
    Hooks in the IRP table Whether hooks exit in the major function table Boolean driverirp
    Malicious attachment Whether there is malware driver attaching to the legitimate device Boolean devicetree
    SSDT hooking Whether there are hooks on SSDT Boolean ssdt
    Abnormal Callbacks Whether there is malicious callback in the system Boolean callbacks
    Abnormal timers Whether there is malicious timer in the system Boolean timers

     | Show Table
    DownLoad: CSV

    In order to improve the classification, we use a vector v to represent the features. The element of the vector consists of two parts: the Boolean value and the weight value. The Boolean value indicates whether a certain feature appears in the memory dump. In our research, there are totally eight features. If the feature shows up in the memory dump, the Boolean value of this feature is true represented by 1, otherwise it is false represented by 0.

    The weight of the feature is gauged by Mutual Information (MI) which measures the statistical dependence of the feature on the class of the sample, as shown in (1).

    I(X;Y)=yYxXp(x,y)log(p(x,y)p(x)p(y)) (1)

    where X represents the frequency of occurrence of a feature in a memory dump and Y is the class of the sample indicating whether or not the sample is a rootkit (i.e. rootkit or benign). p(x) and p(y) are the marginal probability of X and Y. p(x, y) is the joint probability of X and Y. The MI calculated for each feature is used as a weight value, which reduces the noise from less relevant features. The vector is expressed as (2):

    v=((f1,w1),(f2,w2),,(fn1,wn1),(fn,wn)) (2)

    where fn is the Boolean value of the nth feature, and wn is the weight value computed with (1).

    With the feature vector of the memory dumps, machine learning based classifiers can be trained to detect the variants of the malwares and the new emerging ones.

    The machine learning algorithms used to train our classifiers include those based on human perception and bio-inspired models, such as Decision trees, Rule based classifiers, Bayesian and SVM.

    Based on tree structures to make decisions, decision tree is a simple but widely used classification technology. Generally, a decision tree contains one root node, several internal nodes and leaf nodes. Each leaf node is assigned with a class label corresponding to the decision result. The internal node corresponds to an attribute test condition, which is used to distinguish records with different characteristics and divide the sample set into different sub-nodes. The purpose of decision tree learning is to produce a decision tree with strong generalization ability. After building the decision tree, tree-pruning can be carried out to reduce its scale. Therefore, the over fitting problem can be solved and the generalization ability is improved.

    Rule based classifiers uses a set of if…then… rules to classify records. A rule is as follows: f1f2fL. The part on the right side of the symbol is called rule body, which indicates the premise of the rule. The left part is called rule head representing the result of the rule. The quality of classification rules can be gauged with Coverage and Accuracy. Rule sets generated by rule-based classifiers have two important properties: Mutually Exclusive Rule and Exhaustive Rule which are combined to ensure that every record is covered by one and only one rule.

    Bayesian is a method of modeling probabilistic relations between attribute sets and class variables. It combines prior probability of a class with new evidence collected from data. There are two implementations of Bayesian classifier: naive Bayes and Bayesian belief network (BBN). Naive Bayesian classifier assumes conditional independence between attributes when estimating conditional probability of classes. BBN is applicable to problems of classification with certain correlations between attributes.

    SVM is a dual-classification model whose purpose is to find a hyperplane to segment the samples. The principle of segmentation is to maximize the interval, which eventually converts to a convex quadratic programming problem. In the sample space, the partition of hyperplane can be described with the following linear equations: wTx+b=0, in which w is the normal vector which determines the direction of hyperplane, and b is the displacement determining the distance between the hyperplane and origin. It finally comes down to a cubic programming problem, which can be solved with a general quadratic programming algorithm. SMO (Sequential Minimal Optimization) is an efficient algorithm, which can effectively solve the dual problem derived from SVM.

    MalShare [25] and VirusShare [26] are websites of malware repository. In our experiment, around 100 Windows kernel rootkit samples have been collected from them, including FURootkit, NtRootkit etc., as well as their variants for 32bit Windows 7 OS. VirusTotal [27] is an online malware analysis system, through which the type of the collected samples can be confirmed. VirusTotal can also display the variant malwares family.

    Weka [28] is a suite of machine learning software written in Java, containing a collection of tools and algorithms for data analysis and predictive modeling. In our research, Weka 3.9 was used for classifier training. The default file format supported by Weka is Attribute Relation File Format (ARFF). It also supports csv format in which we stored our data.

    To train the classification models, various algorithms supplied by Weka are used, which are divided into four categories: Decision tree, Rule based classifiers, Bayesian and SVM. In each category, the following algorithms are selected: RandomForest, J84 algorithms for Decision tree; JRip, PART algorithms for Rule based classifiers; BayesNet, NaiveBayes for Bayesian classifiers; SMO for SVM. The default configurations of these algorithms are utilized. The machine learning models and corresponding algorithms are shown in Table 3.

    Table 3.  The machine learning models and corresponding algorithms.
    Machine learning models Algorithms
    Decision tree RandomForest, J84
    Rule based classifiers JRip, PART
    Bayesian BayesNet, NaiveBayes
    SVM SMO

     | Show Table
    DownLoad: CSV

    To evaluate the effectiveness of the classifiers trained by the algorithms proposed, we need to answer the following questions:

    ⅰ.  Which classifier we selected has the best performance?

    ⅱ.  Whether or not the scale of the dataset influences the performance of the classifier?

    ⅲ.  Can the trained classifiers effectively detect new rootkits?

    Based on these questions, the following experiments are designed:

    Experiment #1: Ten-fold cross-validation is used to train the chosen classifiers. The extracted features of all the collected samples are used to train the classifiers. There are 12300 records in the dataset, including 10000 malicious and 2300 benign samples. All the classifiers are trained with the default parameters available in Weka. The purpose of this experiment is to verify the best classifier for rootkit detection with the dataset provided.

    Experiment #2: We selected 3000 malicious records and 450 benign records from the dataset of experiment#1 to train the classifiers to verify the influence of dataset scale on the effect of the classifiers.

    Experiment #3: To verify whether our trained classifier can detect unknown malware effectively, we collected some new samples to form a testing dataset, including 500 new malicious samples and 300 benign ones to prove that the classifier can detect new emerging malwares.

    To evaluate the performance of each classifier, we use TPR, FPR, the AUC value (the area under the receiver operating characteristic curve), F-measure and Accuracy as the evaluation criteria. The formulas are shown below, where TP and FP respectively represent the numbers of the correctly classified malwares and benign software, and FN and TN are the numbers of malwares and benign software misidentified. The definitions for these metrics are given as follows.

    TPR=TPTP+FN (3)
    FPR=FPFP+TN (4)
    recall=TPTP+FN (5)
    Precision=TPTN+FP (6)
    Fmeasure = 2PrecisionrecallPrecision+recall (7)
    Accuracy=TP+TNTP+FP+TP+TN (8)

    From the result of experiment #1 shown in Table 4, it can be seen that the Random Forest classifier has the best performance with TRP of 0.996, FRP of 0.079 and AUC of 0.997. The Bayes classifier has the lowest capability for malware detection: the TPR of the NaiveBayes is 0.838, and the FPR and the AUC are 0.630 and 0.922. The JRIP and PART classifier have better performance than SMO.

    Table 4.  The results of experiment #1.
    Classifier TPR FPR AUC F-measure Accuracy
    Random Forest 0.966 0.079 0.997 0.966 0.966
    J48 0.955 0.096 0.990 0.956 0.956
    JRip 0.965 0.079 0.967 0.965 0.965
    PART 0.963 0.079 0.996 0.964 0.964
    BayesNet 0.847 0.630 0.922 0.815 0.827
    NaiveBayes 0.838 0.729 0.950 0.785 0.824
    SMO 0.925 0.016 0.954 0.930 0.947

     | Show Table
    DownLoad: CSV

    Table 5 is the result of experiment #2. Comparing Table 5 with Table 4, the performance of classifiers declines as a whole. For example, the TPR of the Random Forest classifier decreases from 0.966 to 0.946, and FRP increases from 0.079 to 0.096. It indicates that the quantity of the training samples does affect the classifier performance. The classifier trained with mass data would have a better performance.

    Table 5.  The results of experiment #2.
    Classifier TPR FPR AUC F-Measure Accuracy
    Random Forest 0.946 0.096 0.983 0.946 0.946
    J48 0.931 0.100 0.967 0.932 0.934
    JRip 0.903 0.154 0.951 0.904 0.906
    PART 0.953 0.071 0.980 0.953 0.954
    BayesNet 0.776 0.630 0.912 0.801 0.813
    NaiveBayes 0.819 0.603 0.938 0.775 0.825
    SMO 0.935 0.019 0.958 0.938 0.950

     | Show Table
    DownLoad: CSV

    To verify the unknown malware detection ability of the classifier, we collected 500 new malware samples and 300 benign to form a new testing dataset. The classifiers trained in experiment#1 were tested to classify the new samples. From the result of the experiment shown in Table 6, the detection performances are good for most classifiers. It indicates that the classifier can detect unknown malware effectively. The Random Forest classifier has the best performance which can detect unknown kernel rootkits with an Accuracy of 0.986 and an AUC value of 0.998.

    Table 6.  The results of experiment #3.
    Classifier TPR FPR AUC F-measure Accuracy
    Random Forest 0.984 0.076 0.998 0.986 0.986
    J48 0.961 0.084 0.991 0.962 0.965
    JRip 0.967 0.094 0.974 0.971 0.971
    PART 0.969 0.064 0.987 0.966 0.965
    BayesNet 0.856 0.329 0.932 0.891 0.894
    NaiveBayes 0.839 0.583 0.958 0.795 0.864
    SMO 0.965 0.010 0.978 0.958 0.968

     | Show Table
    DownLoad: CSV

    The comparison of TPR, FPR, AUC, F-Measure and Accuracy achieved by all the classifiers in the three experiments is shown in Figures 4-8.

    Figure 4.  The TPR of the three experiments.
    Figure 5.  The FPR of the three experiments.
    Figure 6.  The AUC of the three experiments.
    Figure 7.  The F-Measure of the three experiments.
    Figure 8.  The Accuracy of the three experiments.

    Figures 4 and 5 show the TPR and FPR of the classifiers in the three experiments, respectively. Random Forest has the highest TPR on the whole, which reaches as high as 0.984 in experiment #3. The average value of the TPR of Random Forest in the three experiment is 0.965. The average TPR value of J84, JRip, PART, BayesNet, NaiveBayes and SMO are respectively 0.949, 0.945, 0.962, 0.826, 0.832, 0.942. The classifier of Random Forest has the highest average value of TPR. For FPR, Random forest, J47, PART and SMO seem to produce much lower FPR than BayesNet and NaiveBayes, where SMO has the least FPR among all these approaches.

    Figures 6-8 are respectively the AUC, F-Measure, and Accuracy of the classifiers in the three experiments. As seen, the highest average values of AUC, F-Measure, Accuracy are achieved by the Random Forest classifier, which are 0.992, 0.966 and 0.966. As a result, the Random Forest classifier is the most excellent classifier in our experiment for kernel rootkits detection for VMs.

    The comparison with other related works is summarized in Table 7. For each approach, several classifiers are compared, where the classifier to produce the best results among each group is underlined in the table.

    Table 7.  The Comparison with other related works.
    Related work Classifiers TPR FPR AUC
    Nissim et al. RF, DT, NB, BN, SVM 0.979 0 1
    Cohen et al. RF, BN, AM, LB J48, NB, SMO, Bagging, LR 0.958 0 1
    Bai et al. RF, J48, BN 0.988 0.001 0.998
    Proposed work RF, J48, PART, JRip, NB, BN, SMO 0.984 0.076 0.998
    Random Forest (RF), Decision Trees (DT), Naïve Bayes (NB), Bayesian network (BN), Support Vector Machine (SVM), AdaBoostM1(AM), LogitBoost(LB), Logistic Regression (LR), Partial Decision Tree (PART), where the best classifier within each group is underlined and bold for attention.

     | Show Table
    DownLoad: CSV

    Nissim et al. [16] presented a methodology for trusted detection of unknown malwares for VMs. Experiment #4 in Nissim et al. [16] was aiming at detecting new emerged unknown malwares which is most relevant with our work. Classifiers of Decision Tree, Random Forest, Naïve Bayes, Bayesian network and SVM were trained. The best performance was achieved by Random Forest with the TPR of 0.979, FPR of 0 and AUC of 1.The dataset contained 4000 samples, half of which were benign and the others were malicious. They used 80% of the data for training and 20% for testing.

    Cohen et al. [15] conducted trusted analysis of memory dumps for VMs using Volatility framework, and classifiers of Random Forest, BayesNet, AdaBoostM1 etc. were trained. Experiment 2 in Cohen et al. [15] focused on detecting the unknown infected states which is most relevant with our research. The best performance was reached by Random Forest with the TPR of 0.958, FPR of 0 and AUC of 1. The training dataset consisted of 400 malicious instances and 400 benign ones, the testing dataset consisted of 100 malicious instances and 100 benign ones.

    Bai et al. [21] proposed a malware detection method using ensemble method. Scheme Ⅱ in Bai et al. [21] leveraged multi features to train base classifiers which is most interrelated with our work. Classifiers of Random Forest, J48 and BayesNet were trained on dataset D2. Its best performance was reached by Random Forest with the TPR of 0.988, FRP of 0.001 and AUC of 0.998. The training dataset consisted of 5202 malicious instances and 3918 benign instances, and the testing dataset was composed of 109245 malicious instances and 3953 benign instances.

    The best performance of our proposed work reached as TPR of 0.984, FPR of 0.076 and AUC of 0.998 by Random Forest, which achieved a relatively satisfactory result for VM kernel rootkits detection.

    Aiming at solving the cybersecurity problem of virtual machine in private cloud computing environment, a TKRD method is proposed to detect known and unknown kernel rootkits based on machine learning method. We leveraged memory forensic technology to obtain the malicious tracks in the VM memory. Machine learning technology is used to analyze the malicious behavior of the kernel rootkits and detect malware attacks in the VMs. Experimental results show that our TKRD method can effectively detect kernel rootkits with a best Accuracy of 0.986 and AUC of 0.998.

    The major innovation of our work is summarized as following. First, a trusted way for memory dump acquisition was proposed, where the memory dumps of the VM were collected on the compute nodes at the controller node of the cloud OS OpenStack. For the short and long time running rootkits, the memory capture interval can be adjusted, which can be adaptively set as a tradeoff of efficiency and efficacy. As far as we know, it is the first time that such a method is proposed in VM memory forensics. Second, comprehensive features representing the abnormal behaviors of the kernel rootkits were extracted from the memory dumps, from which we could identify the kernel rootkits effectively. At last, in order to find the classifier with the best performance, various machine learning models were evaluated with the best one i.e. the random forest classifier recommended.

    In future work, deep learning methods [29] such as S-SAE [30] can be used to train malware classifiers. It can not only reduce complexity but also improve the efficacy of feature extraction and the accuracy of data classification. Other classifiers such as SVM-GRBF and CNNs [31] as an extended ANN [32] will be investigated to find the optimal classifying way for malware detection. In addition, we will also explore combined strategies for malware detection as it has shown great potential in Kumara et al. [17], and this will be extended for improved detection of rootkits in future.

    This research was sponsored by the International Research Cooperation Seed Fund of Beijing University of Technology (No. 2018A01).

    The authors declare there is no conflict of interest.


    Acknowledgments



    This research was funded by the program “2do año de continuidad de estancias posdoctorales vinculadas al fortalecimiento de la calidad del posgrado nacional”, granted by Consejo Nacional de Ciencia y Tecnología (CONACYT), Mexico. Grant number I1200/94/2020. M. D Figueroa-Pizano acknowledges CONACyT for the support provided for postdoctoral research.

    Conflict of interest



    All authors declare no conflicts of interest in this paper.

    [1] Asselah T, Durantel D, Pasmant E, et al. (2021) COVID-19: discovery, diagnostics and drug development. J Hepatol 74: 168-184. doi: 10.1016/j.jhep.2020.09.031
    [2] Rojas R, Basto A, Aguilar C, et al. (2018) Prevalencia de diabetes por diagnóstico médico previo en México. Salud Publica Mex 60: 224-232. doi: 10.21149/8566
    [3] Levaillant M, Lièvre G, Baert G (2019) Ending diabetes in Mexico. Lancet 394: 467-468. doi: 10.1016/S0140-6736(19)31662-9
    [4] Singer M (2020) Deadly companions: COVID-19 and diabetes in Mexico. Med Anthropol 39: 660-665. doi: 10.1080/01459740.2020.1805742
    [5] Zhang W, Davis BD, Chen SS, et al. (2021) Emergence of a novel SARS-CoV-2 variant in southern California. JAMA 325: 1324-1326. doi: 10.1001/jama.2021.1612
    [6] Zucman N, Uhel F, Descamps D, et al. (2021) Severe reinfection with South African severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variant 501Y.V2: a case report. Clin Infect Dis ciab129. doi: 10.1093/cid/ciab129
    [7] Fujino T, Nomoto H, Kutsuna S, et al. (2021) Novel SARS-CoV-2 variant in travelers from Brazil to Japan. Emerg Infect Dis 27: 1243-1245. doi: 10.3201/eid2704.210138
    [8] Yadav PD, Nyayanit DA, Sahay RR, et al. (2021) Isolation and characterization of the new SARS-CoV-2 variant in travellers from the United Kingdom to India: VUI-202012/01 of the B.1.1.7 lineage. J Travel Med 28: taab009. doi: 10.1093/jtm/taab009
    [9] Zuckerman NS, Fleishon S, Bucris E, et al. (2021) A unique SARS-CoV-2 spike protein P681H variant detected in Israel. Vaccines 9: 616. doi: 10.3390/vaccines9060616
    [10] Munitz A, Yechezkel M, Dickstein Y, et al. (2021) The rise of SARS-CoV-2 variant B.1.1.7 in Israel intensifies the role of surveillance and vaccination in elderly. medRxiv 2021.
    [11] Mor O, Zuckerman NS, Hazan I, et al. (2021) BNT162b2 vaccination efficacy is marginally affected by the SARS-CoV-2 B.1.351 variant in fully vaccinated individuals. Avaliable at SSRN 3878825 .
    [12] Munitz A, Yechezkel M, Dickstein Y, et al. (2021) Report BNT162b2 vaccination effectively prevents the rapid populations in Israel ll BNT162b2 vaccination effectively prevents the rapid rise of SARS-CoV-2 variant B.1.1.7 in high-risk populations in Israel. Cell Rep Med 2: 100264. doi: 10.1016/j.xcrm.2021.100264
    [13] Haas EJ, Angulo FJ, Mclaughlin JM, et al. (2021) Impact and effectiveness of mRNA BNT162b2 vaccine against SARS-CoV-2 infections and COVID-19 cases, hospitalisations, and deaths following a nationwide vaccination campaign in Israel: an observational study using national surveillance data. Lancet 397: 1819-1829. doi: 10.1016/S0140-6736(21)00947-8
    [14] Kustin T, Harel N, Finkel U, et al. (2021) Evidence for increased breakthrough rates of SARS-CoV-2 variants of concern in BNT162b2-mRNA-vaccinated individuals. Nat Med 27: 1379-1384. doi: 10.1038/s41591-021-01413-7
    [15] Hu T, Liu Y, Zhao M, et al. (2020) A comparison of COVID-19, SARS and MERS. PeerJ 8: e9725. doi: 10.7717/peerj.9725
    [16] Harrison AG, Lin T, Wang P (2020) Mechanisms of SARS-CoV-2 transmission and pathogenesis. Trends Immunol 41: 1100-1115. doi: 10.1016/j.it.2020.10.004
    [17] Yuki K, Fujiogi M, Koutsogiannaki S (2020) COVID-19 pathophysiology: a review. Clin Immunol 215: 108427. doi: 10.1016/j.clim.2020.108427
    [18] Sardu C, Gambardella J, Morelli MB, et al. (2020) Hypertension, thrombosis, kidney failure, and diabetes: is COVID-19 an endothelial disease? A comprehensive evaluation of clinical and basic evidence. J Clin Med 9: 1417. doi: 10.3390/jcm9051417
    [19] Safwat MA, Abdul-Rahman A, Forsan H, et al. (2021) COVID-19; Immunology, pathology, severity and immunosuppressants. Azhar Int J Pharm Med Sci 1: 1-14. doi: 10.21608/aijpms.2021.50103.1005
    [20] Kabashneh S, Ali H, Alkassis S (2020) Multi-Organ failure in a patient with diabetes due to COVID-19 with clear lungs. Cureus 12: e8147.
    [21] Guo L, Shi Z, Zhang Y, et al. (2020) Comorbid diabetes and the risk of disease severity or death among 8807 COVID-19 patients in China: a meta-analysis. Diabetes Res Clin Pract 166: 108346. doi: 10.1016/j.diabres.2020.108346
    [22] Zhou Y, Chi J, Lv W, et al. (2021) Obesity and diabetes as high-risk factors for severe coronavirus disease 2019 (Covid-19). Diabetes Metab Res Rev 37: e3377.
    [23] Parveen R, Sehar N, Bajpai R, et al. (2020) Association of diabetes and hypertension with disease severity in covid-19 patients: a systematic literature review and exploratory meta-analysis. Diabetes Res Clin Pract 166: 108295. doi: 10.1016/j.diabres.2020.108295
    [24] Frydrych LM, Bian G, O'Lone DE, et al. (2018) Obesity and type 2 diabetes mellitus drive immune dysfunction, infection development, and sepsis mortality. J Leukoc Biol 104: 525-534. doi: 10.1002/JLB.5VMR0118-021RR
    [25] Sur J, Sharma J, Sharma D (2021) Diabetes might augment the severity of COVID-19: a current prospects. Front Cardiovasc Med 7: 613255. doi: 10.3389/fcvm.2020.613255
    [26] Kumar A, Arora A, Sharma P, et al. (2020) Is diabetes mellitus associated with mortality and severity of COVID-19? A meta-analysis. Diabetes Metab Syndr 14: 535-545. doi: 10.1016/j.dsx.2020.04.044
    [27] Jeong IK, Yoon KH, Lee MK (2020) Diabetes and COVID-19: global and regional perspectives. Diabetes Res Clin Pract 166: 108303. doi: 10.1016/j.diabres.2020.108303
    [28] Chen Y, Yang D, Cheng B, et al. (2020) Clinical characteristics and outcomes of patients with diabetes and COVID-19 in association with glucose-lowering medication. Diabetes Care 43: 1399-1407. doi: 10.2337/dc20-0660
    [29] Mantovani A, Byrne CD, Zheng MH, et al. (2020) Diabetes as a risk factor for greater COVID-19 severity and in-hospital death: a meta-analysis of observational studies. Nutr Metab Cardiovasc Dis 30: 1236-1248. doi: 10.1016/j.numecd.2020.05.014
    [30] Du M, Lin YX, Yan WX, et al. (2020) Prevalence and impact of diabetes in patients with COVID-19 in China. World J Diabetes 11: 468-480. doi: 10.4239/wjd.v11.i10.468
    [31] Targher G, Mantovani A, Wang XB, et al. (2020) Patients with diabetes are at higher risk for severe illness from COVID-19. Diabetes Metab 46: 335-337. doi: 10.1016/j.diabet.2020.05.001
    [32] Ciardullo S, Zerbini F, Perra S, et al. (2020) Impact of diabetes on COVID-19-related in-hospital mortality: a retrospective study from Northern Italy. J Endocrinol Invest 44: 843-850. doi: 10.1007/s40618-020-01382-7
    [33] Halvatsiotis P, Kotanidou A, Tzannis K, et al. (2020) Demographic and clinical features of critically ill patients with COVID-19 in Greece: the burden of diabetes and obesity. Diabetes Res Clin Pract 166: 108331. doi: 10.1016/j.diabres.2020.108331
    [34] Nachtigall I, Lenga P, Jóźwiak K, et al. (2020) Clinical course and factors associated with outcomes among 1904 patients hospitalized with COVID-19 in Germany: an observational study. Clin Microbiol Infect 26: 1663-1669. doi: 10.1016/j.cmi.2020.08.011
    [35] Sutter W, Duceau B, Vignac M, et al. (2020) Association of diabetes and outcomes in patients with COVID-19: propensity score-matched analyses from a French retrospective cohort. Diabetes Metab 47: 101222. doi: 10.1016/j.diabet.2020.101222
    [36] Barron E, Bakhai C, Kar P, et al. (2020) Associations of type 1 and type 2 diabetes with COVID-19-related mortality in England: a whole-population study. Lancet Diabetes Endocrinol 8: 813-822. doi: 10.1016/S2213-8587(20)30272-2
    [37] Ge Y, Sun S, Shen Y (2021) Estimation of case-fatality rate in COVID-19 patients with hypertension and diabetes mellitus in the New York State: a preliminary report. Epidemiol Infect 149: e14. doi: 10.1017/S0950268821000066
    [38] Hussain S, Baxi H, Chand Jamali M, et al. (2020) Burden of diabetes mellitus and its impact on COVID-19 patients: a meta-analysis of real-world evidence. Diabetes Metab Syndr 14: 1595-1602. doi: 10.1016/j.dsx.2020.08.014
    [39] Ashktorab H, Pizuorno A, Oskroch G, et al. (2021) COVID-19 in Latin America: symptoms, morbidities, and gastrointestinal manifestations. Gastroenterology 160: 938-940. doi: 10.1053/j.gastro.2020.10.033
    [40] Izzi-Engbeaya C, Distaso W, Amin A, et al. (2021) Adverse outcomes in COVID-19 and diabetes: a retrospective cohort study from three London teaching hospitals. BMJ Open Diabetes Res Care 9: e001858. doi: 10.1136/bmjdrc-2020-001858
    [41] Ebekozien O, Agarwal S, Noor N, et al. (2020) Inequities in diabetic ketoacidosis among patients with type 1 diabetes and COVID-19: data from 52 US clinical centers. J Clin Endocrinol Metab 106: e1755-e1762. doi: 10.1210/clinem/dgaa920
    [42] Hafidh K, Abbas S, Khan A, et al. (2020) The clinical characteristics and outcomes of COVID-19 infections in patients with diabetes at a tertiary care center in the UAE. Dubai Diabetes Endocrinol J 26: 158-163. doi: 10.1159/000512232
    [43] Yan Y, Yang Y, Wang F, et al. (2020) Clinical characteristics and outcomes of patients with severe covid-19 with diabetes. BMJ Open Diabetes Res Care 8: e001343. doi: 10.1136/bmjdrc-2020-001343
    [44] Ling P, Luo S, Zheng X, et al. (2021) Elevated fasting blood glucose within the first week of hospitalization was associated with progression to severe illness of COVID-19 in patients with preexisting diabetes: a multicenter observational study. J Diabetes 13: 89-93. doi: 10.1111/1753-0407.13121
    [45] Silverio A, Di Maio M, Citro R, et al. (2021) Cardiovascular risk factors and mortality in hospitalized patients with COVID-19: systematic review and meta-analysis of 45 studies and 18,300 patients. BMC Cardiovasc Disord 21: 23. doi: 10.1186/s12872-020-01816-3
    [46] Acharya D, Lee K, Lee DS, et al. (2020) Mortality rate and predictors of mortality in hospitalized COVID-19 patients with diabetes. Healthcare 8: 338. doi: 10.3390/healthcare8030338
    [47] Bode B, Garrett V, Messler J, et al. (2020) Glycemic characteristics and clinical outcomes of COVID-19 patients hospitalized in the United States. J Diabetes Sci Technol 14: 813-821. doi: 10.1177/1932296820924469
    [48] Bhandari S, Rankawat G, Singh A, et al. (2020) Impact of glycemic control in diabetes mellitus on management of COVID-19 infection. Int J Diabetes Dev Ctries 1-6.
    [49] Ceriello A, De Nigris V, Prattichizzo F (2020) Why is hyperglycaemia worsening COVID-19 and its prognosis? Diabetes Obes Metab 22: 1951-1952. doi: 10.1111/dom.14098
    [50] Lee MH, Wong C, Ng CH, et al. (2021) Effects of hyperglycaemia on complications of COVID-19: a meta-analysis of observational studies. Diabetes Obes Metab 23: 287-289. doi: 10.1111/dom.14184
    [51] Singh AK, Singh R (2020) Hyperglycemia without diabetes and new-onset diabetes are both associated with poorer outcomes in COVID-19. Diabetes Res Clin Pract 167: 108382. doi: 10.1016/j.diabres.2020.108382
    [52] Carrasco-Sánchez FJ, López-Carmona MD, Martínez-Marcos FJ, et al. (2021) Admission hyperglycaemia as a predictor of mortality in patients hospitalized with COVID-19 regardless of diabetes status: data from the Spanish SEMI-COVID-19 Registry. Ann Med 53: 103-116. doi: 10.1080/07853890.2020.1836566
    [53] Wang S, Ma P, Zhang S, et al. (2020) Fasting blood glucose at admission is an independent predictor for 28-day mortality in patients with COVID-19 without previous diagnosis of diabetes: a multi-centre retrospective study. Diabetologia 63: 2102-2111. doi: 10.1007/s00125-020-05209-1
    [54] Li H, Tian S, Chen T, et al. (2020) Newly diagnosed diabetes is associated with a higher risk of mortality than known diabetes in hospitalized patients with COVID-19. Diabetes Obes Metab 22: 1897-1906. doi: 10.1111/dom.14099
    [55] Lim S, Bae JH, Kwon HS, et al. (2021) COVID-19 and diabetes mellitus: from pathophysiology to clinical management. Nat Rev Endocrinol 17: 11-30. doi: 10.1038/s41574-020-00435-4
    [56] Abu-Farha M, Al-Mulla F, Thanaraj TA, et al. (2020) Impact of diabetes in patients diagnosed with COVID-19. Front Immunol 11: 576818. doi: 10.3389/fimmu.2020.576818
    [57] Yaribeygi H, Sathyapalan T, Jamialahmadi T, et al. (2020) The impact of diabetes mellitus in COVID-19: a mechanistic review of molecular interactions. J Diabetes Res 2020: 5436832.
    [58] Singh AK, Gupta R, Ghosh A, et al. (2020) Diabetes in COVID-19: prevalence, pathophysiology, prognosis and practical considerations. Diabetes Metab Syndr 14: 303-310. doi: 10.1016/j.dsx.2020.04.004
    [59] Erener S (2020) Diabetes, infection risk and COVID-19. Mol Metab 39: 101044. doi: 10.1016/j.molmet.2020.101044
    [60] Azar WS, Njeim R, Fares AH, et al. (2020) COVID-19 and diabetes mellitus: how one pandemic worsens the other. Rev Endocr Metab Disord 21: 451-463. doi: 10.1007/s11154-020-09573-6
    [61] Banik GR, Alqahtani AS, Booy R, et al. (2016) Risk factors for severity and mortality in patients with MERS-CoV: analysis of publicly available data from Saudi Arabia. Virol Sin 31: 81-84. doi: 10.1007/s12250-015-3679-z
    [62] Kulcsar KA, Coleman CM, Beck SE, et al. (2019) Comorbid diabetes results in immune dysregulation and enhanced disease severity following MERS-CoV infection. JCI Insight 4: e131774. doi: 10.1172/jci.insight.131774
    [63] Halaji M, Farahani A, Ranjbar R, et al. (2020) Emerging coronaviruses: first SARS, second MERS and third SARS-COV-2: epidemiological updates of COVID-19. Infez Med 28: 6-17.
    [64] Muniyappa R, Gubbi S (2020) COVID-19 pandemic, coronaviruses, and diabetes mellitus. Am J Physiol Endocrinol Metab 318: E736-E741. doi: 10.1152/ajpendo.00124.2020
    [65] Petersmann A, Nauck M, Müller-Wieland D, et al. (2018) Definition, classification and diagnosis of diabetes mellitus. Exp Clin Endocrinol Diabetes 127: S1-S7.
    [66] Punthakee Z, Goldenberg R, et al. (2018) Definition, classification and diagnosis of diabetes, prediabetes and metabolic syndrome. Can J Diabetes 42: S10-S15. doi: 10.1016/j.jcjd.2017.10.003
    [67] Daryabor G, Atashzar MR, Kabelitz D, et al. (2020) The effects of type 2 diabetes mellitus on organ metabolism and the immune system. Front Immunol 11: 1582. doi: 10.3389/fimmu.2020.01582
    [68] Papatheodorou K, Banach M, Bekiari E, et al. (2018) Complications of diabetes 2017. J Diabetes Res 2018: 3086167. doi: 10.1155/2018/3086167
    [69] Huang D, Refaat M, Mohammedi K, et al. (2017) Macrovascular complications in patients with diabetes and prediabetes. Biomed Res Int 2017: 7839101.
    [70] Khalil H (2017) Diabetes microvascular complications – A clinical update. Diabetes Metab Syndr 11: S133-S139. doi: 10.1016/j.dsx.2016.12.022
    [71] Struijs JN, Baan CA, Schellevis FG, et al. (2006) Comorbidity in patients with diabetes mellitus: impact on medical health care utilization. BMC Health Serv Res 6: 84. doi: 10.1186/1472-6963-6-84
    [72] Prattichizzo F, de Candia P, Nicolucci A, et al. (2021) Elevated HbA1c levels in pre-Covid-19 infection increases the risk of mortality: a sistematic review and meta-analysis. Diabetes Metab Res Rev e3476.
    [73] Mirani M, Favacchio G, Carrone F, et al. (2020) Impact of comorbidities and glycemia at admission and dipeptidyl peptidase 4 inhibitors in patients with type 2 diabetes with covid-19: a case series from an academic hospital in lombardy, Italy. Diabetes Care 43: 3042-3049. doi: 10.2337/dc20-1340
    [74] Orioli L, Servais T, Belkhir L, et al. (2021) Clinical characteristics and short-term prognosis of in-patients with diabetes and COVID-19: a retrospective study from an academic center in Belgium. Diabetes Metab Syndr 15: 149-157. doi: 10.1016/j.dsx.2020.12.020
    [75] Sun Y, Guan X, Jia L, et al. (2021) Independent and combined effects of hypertension and diabetes on clinical outcomes in patients with COVID-19: a retrospective cohort study of Huoshen Mountain Hospital and Guanggu Fangcang Shelter Hospital. J Clin Hypertens 23: 218-231. doi: 10.1111/jch.14146
    [76] Vieira E, Mirizio GG, Barin GR, et al. (2020) Clock genes, inflammation and the immune system—Implications for diabetes, obesity and neurodegenerative diseases. Int J Mol Sci 21: 9743. doi: 10.3390/ijms21249743
    [77] Zhou T, Hu Z, Yang S, et al. (2018) Role of adaptive and innate immunity in type 2 diabetes mellitus. J Diabetes Res 2018: 7457269.
    [78] Hodgson K, Morris J, Bridson T, et al. (2015) Immunological mechanisms contributing to the double burden of diabetes and intracellular bacterial infections. Immunology 144: 171-185. doi: 10.1111/imm.12394
    [79] Alzaid F, Julla J, Diedisheim M, et al. (2020) Monocytopenia, monocyte morphological anomalies and hyperinflammation characterise severe COVID-19 in type 2 diabetes. EMBO Mol Med 12: e13038. doi: 10.15252/emmm.202013038
    [80] De Candia P, Prattichizzo F, Garavelli S, et al. (2019) Type 2 diabetes: how much of an autoimmune disease? Front Endocrinol (Lausanne) 10: 451. doi: 10.3389/fendo.2019.00451
    [81] Prattichizzo F, Giuliani A, Sabbatinelli J, et al. (2020) Prevalence of residual inflammatory risk and associated clinical variables in patients with type 2 diabetes. Diabetes Obes Metab 22: 1696-1700. doi: 10.1111/dom.14081
    [82] Maedler K, Spinas GA, Lehmann R, et al. (2001) Glucose induces β-Cell apoptosis via upregulation of the Fas receptor in human islets. Diabetes 50: 1683-1690. doi: 10.2337/diabetes.50.8.1683
    [83] Maedler K, Sergeev P, Ris F, et al. (2002) Glucose-induced β cell production of IL-1β contributes to glucotoxicity in human pancreatic islets. J Clin Invest 110: 851-860. doi: 10.1172/JCI200215318
    [84] Zhou W, Ye S, Wang W, et al. (2020) Clinical features of COVID-19 patients with diabetes and secondary hyperglycemia. J Diabetes Res 2020: 3918723.
    [85] Roganović J (2021) Downregulation of microRNA-146a in diabetes, obesity and hypertension may contribute to severe COVID-19. Med Hypotheses 146: 110448. doi: 10.1016/j.mehy.2020.110448
    [86] Gou L, Xiang M, Ran X, et al. (2021) Hyperosmolarity deserves more attention in critically Ill COVID-19 patients with diabetes: a cohort-based study. Diabetes Metab Syndr Obes Targets Ther 14: 47-58. doi: 10.2147/DMSO.S284148
    [87] Zheng M, Wang X, Guo H, et al. (2021) The cytokine profiles and immune response are increased in COVID-19 patients with type 2 diabetes mellitus. J Diabetes Res 2021: 9526701. doi: 10.1155/2021/9526701
    [88] Guo W, Li M, Dong Y, et al. (2020) Diabetes is a risk factor for the progression and prognosis of COVID-19. Diabetes Metab Res Rev e3319.
    [89] Cheng Y, Yue L, Wang Z, et al. (2021) Hyperglycemia associated with lymphopenia and disease severity of COVID-19 in type 2 diabetes mellitus. J Diabetes Complications 35: 107809. doi: 10.1016/j.jdiacomp.2020.107809
    [90] Sun Y, Zhao R, Hu Z, et al. (2020) Differences in the clinical and hematological characteristics of COVID-19 patients with and without type 2 diabetes. J Diabetes Res 2020: 1038585.
    [91] Zhang Y, Li H, Zhang J, et al. (2020) The clinical characteristics and outcomes of patients with diabetes and secondary hyperglycaemia with coronavirus disease 2019: a single-centre, retrospective, observational study in Wuhan. Diabetes Obes Metab 22: 1443-1454. doi: 10.1111/dom.14086
    [92] Ulhaq ZS, Soraya GV (2020) Interleukin-6 as a potential biomarker of COVID-19 progression. Médecine Mal Infect 50: 382-383. doi: 10.1016/j.medmal.2020.04.002
    [93] Scheller J, Chalaris A, Schmidt-Arras D, et al. (2011) The pro- and anti-inflammatory properties of the cytokine interleukin-6. Biochim Biophys Acta 1813: 878-888. doi: 10.1016/j.bbamcr.2011.01.034
    [94] Muangchan C, Pope JE (2013) The significance of interleukin-6 and C-reactive protein in systemic sclerosis: a systematic literature review. Clin Exp Rheumatol 31: 122-134.
    [95] Mishra Y, Pathak BK, Mohakuda SS, et al. (2020) Relation of D-dimer levels of COVID-19 patients with diabetes mellitus. Diabetes Metab Syndr 14: 1927-1930. doi: 10.1016/j.dsx.2020.09.035
    [96] Rostami M, Mansouritorghabeh H (2020) D-dimer level in COVID-19 infection: a systematic review. Expert Rev Hematol 13: 1265-1275. doi: 10.1080/17474086.2020.1831383
    [97] Mei H, Luo L, Hu Y (2020) Thrombocytopenia and thrombosis in hospitalized patients with COVID-19. J Hematol Oncol 13: 161. doi: 10.1186/s13045-020-01003-z
    [98] Zahran AM, El-Badawy O, Mohamad IL, et al. (2018) Platelet activation and platelet–leukocyte aggregates in type I diabetes mellitus. Clin Appl Thromb Hemost 24: 230S-239S. doi: 10.1177/1076029618805861
    [99] Kaur R, Kaur M, Singh J (2018) Endothelial dysfunction and platelet hyperactivity in type 2 diabetes mellitus: molecular insights and therapeutic strategies. Cardiovasc Diabetol 17: 121. doi: 10.1186/s12933-018-0763-3
    [100] Shang J, Wang Q, Zhang H, et al. (2021) The relationship between diabetes mellitus and COVID-19 prognosis: a retrospective cohort study in Wuhan, China. Am J Med 134: e6-e14. doi: 10.1016/j.amjmed.2020.05.033
    [101] Guo T, Shen Q, Ouyang X, et al. (2021) Clinical findings in diabetes mellitus patients with COVID-19. J Diabetes Res 2021: 7830136.
    [102] Al-Salameh A, Lanoix JP, Bennis Y, et al. (2021) Characteristics and outcomes of COVID-19 in hospitalized patients with and without diabetes. Diabetes Metab Res Rev 37: e3388. doi: 10.1002/dmrr.3388
    [103] Chen X, Hu W, Ling J, et al. (2020) Hypertension and diabetes delay the viral clearance in COVID-19 patients. medRxiv .
    [104] Clarke NE, Turner AJ (2012) Angiotensin-converting enzyme 2: the first decade. Int J Hypertens 2012: 307315. doi: 10.1155/2012/307315
    [105] Warner FJ, Smith AI, Hooper NM, et al. (2004) Angiotensin-converting enzyme-2: a molecular and cellular perspective. Cell Mol Life Sci 61: 2704-2713. doi: 10.1007/s00018-004-4240-7
    [106] Santos RAS, Ferreira AJ, Simões E, Silva AC (2008) Recent advances in the angiotensin-converting enzyme 2-angiotensin(1-7)-Mas axis. Exp Physiol 93: 519-527. doi: 10.1113/expphysiol.2008.042002
    [107] Santos RAS, Ferreira AJ, Verano-Braga T, et al. (2013) Angiotensin-converting enzyme 2, angiotensin-(1-7) and Mas: new players of the renin-angiotensin system. J Endocrinol 216: R1-R17. doi: 10.1530/JOE-12-0341
    [108] Hashimoto T, Perlot T, Rehman A, et al. (2012) ACE2 links amino acid malnutrition to microbial ecology and intestinal inflammation. Nature 487: 477-481. doi: 10.1038/nature11228
    [109] Perlot T, Penninger JM (2013) ACE2 – From the renin-angiotensin system to gut microbiota and malnutrition. Microbes Infect 15: 866-873. doi: 10.1016/j.micinf.2013.08.003
    [110] Hamming I, Timens W, Bulthuis MLC, et al. (2004) Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis. J Pathol 203: 631-637. doi: 10.1002/path.1570
    [111] Tikellis C, Johnston CI, Forbes JM, et al. (2003) Characterization of renal angiotensin-converting enzyme 2 in diabetic nephropathy. Hypertension 41: 392-397. doi: 10.1161/01.HYP.0000060689.38912.CB
    [112] Xia H, Lazartigues E (2008) Angiotensin-converting enzyme 2 in the brain: properties and future directions. J Neurochem 107: 1482-1494. doi: 10.1111/j.1471-4159.2008.05723.x
    [113] Xu H, Zhong L, Deng J, et al. (2020) High expression of ACE2 receptor of 2019-nCoV on the epithelial cells of oral mucosa. Int J Oral Sci 12: 8. doi: 10.1038/s41368-020-0074-x
    [114] To KF, Lo AWI (2004) Exploring the pathogenesis of severe acute respiratory syndrome (SARS): The tissue distribution of the coronavirus (SARS-CoV) and its putative receptor, angiotensin-converting enzyme 2 (ACE2). J Pathol 203: 740-743. doi: 10.1002/path.1597
    [115] Batlle D, Wysocki J, Satchell K (2020) Soluble angiotensin-converting enzyme 2: a potential approach for coronavirus infection therapy? Clin Sci 134: 543-545. doi: 10.1042/CS20200163
    [116] Ni W, Yang X, Yang D, et al. (2020) Role of angiotensin-converting enzyme 2 (ACE2) in COVID-19. Crit Care 24: 422. doi: 10.1186/s13054-020-03120-0
    [117] Zhang H, Penninger JM, Li Y, et al. (2020) Angiotensin-converting enzyme 2 (ACE2) as a SARS-CoV-2 receptor: molecular mechanisms and potential therapeutic target. Intensive Care Med 46: 586-590. doi: 10.1007/s00134-020-05985-9
    [118] Shang J, Wan Y, Luo C, et al. (2020) Cell entry mechanisms of SARS-CoV-2. Proc Natl Acad Sci U S A 117: 11727-11734. doi: 10.1073/pnas.2003138117
    [119] Valencia I, Peiró C, Lorenzo Ó, et al. (2020) DPP4 and ACE2 in diabetes and COVID-19: therapeutic targets for cardiovascular complications? Front Pharmacol 11: 1161. doi: 10.3389/fphar.2020.01161
    [120] Bourgonje AR, Abdulle AE, Timens W, et al. (2020) Angiotensin-converting enzyme 2 (ACE2), SARS-CoV-2 and the pathophysiology of coronavirus disease 2019 (COVID-19). J Pathol 251: 228-248. doi: 10.1002/path.5471
    [121] Chen J, Jiang Q, Xia X, et al. (2020) Individual variation of the SARS-CoV-2 receptor ACE2 gene expression and regulation. Aging Cell 19: e13168.
    [122] Mizuiri S, Hemmi H, Arita M, et al. (2008) Expression of ACE and ACE2 in individuals with diabetic kidney disease and healthy controls. Am J Kidney Dis 51: 613-623. doi: 10.1053/j.ajkd.2007.11.022
    [123] Reich HN, Oudit GY, Penninger JM, et al. (2008) Decreased glomerular and tubular expression of ACE2 in patients with type 2 diabetes and kidney disease. Kidney Int 74: 1610-1616. doi: 10.1038/ki.2008.497
    [124] Gilbert RE, Caldwell L, Misra PS, et al. (2021) Overexpression of the severe acute respiratory syndrome coronavirus-2 receptor, angiotensin-converting enzyme 2, in diabetic kidney disease: implications for kidney injury in novel coronavirus disease 2019. Can J Diabetes 45: 162-166.e1. doi: 10.1016/j.jcjd.2020.07.003
    [125] Soldo J, Heni M, Königsrainer A, et al. (2020) Increased hepatic ace2 expression in nafl and diabetes-a risk for covid-19 patients? Diabetes Care 43: e134-e136. doi: 10.2337/dc20-1458
    [126] Wijnant SRA, Jacobs M, Van Eeckhoutte HP, et al. (2020) Expression of ACE2, the SARS-CoV-2 receptor, in lung tissue of patients with type 2 diabetes. Diabetes 69: 2691-2699. doi: 10.2337/db20-0669
    [127] Parit R, Jayavel S (2021) Association of ACE inhibitors and angiotensin type II blockers with ACE2 overexpression in COVID-19 comorbidities: a pathway-based analytical study. Eur J Pharmacol 896: 173899. doi: 10.1016/j.ejphar.2021.173899
    [128] Pal R, Bhadada SK (2020) COVID-19 and diabetes mellitus: an unholy interaction of two pandemics. Diabetes Metab Syndr 14: 513-517. doi: 10.1016/j.dsx.2020.04.049
    [129] Menon R, Otto EA, Sealfon R, et al. (2020) SARS-CoV-2 receptor networks in diabetic and COVID-19–associated kidney disease. Kidney Int 98: 1502-1518. doi: 10.1016/j.kint.2020.09.015
    [130] Batchu SN, Kaur H, Yerra VG, et al. (2021) Lung and kidney ACE2 and TMPRSS2 in renin-angiotensin system blocker-treated comorbid diabetic mice mimicking host factors that have been linked to severe COVID-19. Diabetes 70: 759-771. doi: 10.2337/db20-0765
    [131] Sriram K, Insel PA (2020) Risks of ACE inhibitor and ARB usage in COVID-19: evaluating the evidence. Clin Pharmacol Ther 108: 236-241. doi: 10.1002/cpt.1863
    [132] Burrell LM, Harrap SB, Velkoska E, et al. (2013) The ACE2 gene: its potential as a functional candidate for cardiovascular disease. Clin Sci 124: 65-76. doi: 10.1042/CS20120269
    [133] Meng N, Zhang Y, Ma J, et al. (2015) Association of polymorphisms of angiotensin i converting enzyme 2 with retinopathy in type 2 diabetes mellitus among Chinese individuals. Eye 29: 266-271. doi: 10.1038/eye.2014.254
    [134] Fröjdö S, Sjölind L, Parkkonen M, et al. (2005) Polymorphisms in the gene encoding angiotensin I converting enzyme 2 and diabetic nephropathy. Diabetologia 48: 2278-2281. doi: 10.1007/s00125-005-1955-4
    [135] Currie D, McKnight AJ, Patterson CC, et al. (2010) Investigation of ACE, ACE2 and AGTR1 genes for association with nephropathy in type 1 diabetes mellitus. Diabet Med 27: 1188-1194. doi: 10.1111/j.1464-5491.2010.03097.x
    [136] Chaoxin J, Daili S, Yanxin H, et al. (2013) The influence of angiotensin-converting enzyme 2 gene polymorphisms on type 2 diabetes mellitus and coronary heart disease. Eur Rev Med Pharmacol Sci 17: 2654-2659.
    [137] Patel SK, Wai B, Ord M, et al. (2012) Association of ACE2 genetic variants with blood pressure, left ventricular mass, and cardiac function in caucasians with type 2 diabetes. Am J Hypertens 25: 216-222. doi: 10.1038/ajh.2011.188
    [138] Sayed S (2021) COVID-19 and diabetes; Possible role of polymorphism and rise of telemedicine. Prim Care Diabetes 15: 4-9. doi: 10.1016/j.pcd.2020.08.018
    [139] Calcagnile M, Forgez P, Iannelli A, et al. (2021) Molecular docking simulation reveals ACE2 polymorphisms that may increase the affinity of ACE2 with the SARS-CoV-2 spike protein. Biochimie 180: 143-148. doi: 10.1016/j.biochi.2020.11.004
    [140] Calcagnile M, Forgez P, Iannelli A, et al. (2020) ACE2 polymorphisms and individual susceptibility to SARS-CoV-2 infection: insights from an in silico study. bioRxiv .
    [141] Hamet P, Pausova Z, Attaoua R, et al. (2021) SARS-COV-2 receptor ACE2 gene is associated with hypertension and severity of COVID 19: interaction with sex, obesity and smoking. Am J Hypertens 34: 367-376. doi: 10.1093/ajh/hpaa223
    [142] Choudhary S, Sreenivasulu K, Mitra P, et al. (2020) Role of genetic variants and gene expression in the susceptibility and severity of COVID-19. Ann Lab Med 41: 129-138. doi: 10.3343/alm.2021.41.2.129
    [143] Paniri A, Hosseini MM, Moballegh-Eslam M, et al. (2021) Comprehensive in silico identification of impacts of ACE2 SNPs on COVID-19 susceptibility in different populations. Gene Rep 22: 100979. doi: 10.1016/j.genrep.2020.100979
    [144] Wee AKH (2021) COVID-19's toll on the elderly and those with diabetes mellitus – Is vitamin B12 deficiency an accomplice? Med Hypotheses 146: 110374. doi: 10.1016/j.mehy.2020.110374
    [145] Gonçalves SEAB, Gonçalves TJM, Guarnieri A, et al. (2021) Association between thiamine deficiency and hyperlactatemia among critically ill patients with diabetes infected by SARS-CoV-2. J Diabetes 13: 413-419. doi: 10.1111/1753-0407.13156
    [146] Rhodes JM, Subramanian S, Laird E, et al. (2021) Perspective: vitamin D deficiency and COVID-19 severity – plausibly linked by latitude, ethnicity, impacts on cytokines, ACE2 and thrombosis. J Intern Med 289: 97-115. doi: 10.1111/joim.13149
    [147] Tort AR, Montelongo EAM, Cuazitl AM, et al. (2020) La deficiencia de vitamina D es un factor de riesgo de mortalidad en pacientes con COVID-19. Rev Sanid Milit 74: 106-113. doi: 10.35366/93773
    [148] Biesalski HK (2020) Vitamin D deficiency and co-morbidities in COVID-19 patients – A fatal relationship? NFS J 20: 10. doi: 10.1016/j.nfs.2020.06.001
    [149] Berridge MJ (2017) Vitamin D deficiency and diabetes. Biochem J 474: 1321-1332. doi: 10.1042/BCJ20170042
    [150] Luo BA, Gao F, Qin LL (2017) The association between vitamin D deficiency and diabetic retinopathy in type 2 diabetes: A meta-analysis of observational studies. Nutrients 9: 307. doi: 10.3390/nu9030307
    [151] Singh SK, Jain R, Singh S (2020) Vitamin D deficiency in patients with diabetes and COVID-19 infection. Diabetes Metab Syndr 14: 1033-1035. doi: 10.1016/j.dsx.2020.06.071
    [152] Jin S, Hu W (2021) Severity of COVID-19 and treatment strategy for patient with diabetes. Front Endocrinol (Lausanne) 12: 602735. doi: 10.3389/fendo.2021.602735
    [153] Ceriello A, Prattichizzo F (2021) Pharmacological management of COVID-19 in type 2 diabetes. J Diabetes Complications 35: 107927. doi: 10.1016/j.jdiacomp.2021.107927
    [154] Deng F, Gao D, Ma X, et al. (2021) Corticosteroids in diabetes patients infected with COVID-19. Ir J Med Sci 190: 29-31. doi: 10.1007/s11845-020-02287-3
    [155] Stoian AP, Catrinoiu D, Rizzo M, et al. (2021) Hydroxychloroquine, COVID-19 and diabetes. Why it is a different story. Diabetes Metab Res Rev 37: e3379. doi: 10.1002/dmrr.3379
    [156] Khunti K, Knighton P, Zaccardi F, et al. (2021) Prescription of glucose-lowering therapies and risk of COVID-19 mortality in people with type 2 diabetes : a nationwide observational study in England. Lancet Diabetes Endocrinol 9: 293-303. doi: 10.1016/S2213-8587(21)00050-4
    [157] Gregory JM, Slaughter JC, Duffus SH, et al. (2021) COVID-19 severity is tripled in the diabetes community: a prospective analysis of the pandemic's impact in type 1 and type 2 diabetes. Diabetes Care 44: 526-532. doi: 10.2337/dc20-2260
    [158] Fisher L, Polonsky W, Asuni A, et al. (2020) The early impact of the COVID-19 pandemic on adults with type 1 or type 2 diabetes: a national cohort study. J Diabetes Complications 34: 107748. doi: 10.1016/j.jdiacomp.2020.107748
    [159] Ruissen MM, Regeer H, Landstra CP, et al. (2021) Increased stress, weight gain and less exercise in relation to glycemic control in people with type 1 and type 2 diabetes during the COVID-19 pandemic. BMJ Open Diabetes Res Care 9: e002035. doi: 10.1136/bmjdrc-2020-002035
    [160] Holman N, Knighton P, Kar P, et al. (2020) Risk factors for COVID-19-related mortality in people with type 1 and type 2 diabetes in England: a population-based cohort study. Lancet Diabetes Endocrinol 8: 823-833. doi: 10.1016/S2213-8587(20)30271-0
    [161] Beliard K, Ebekozien O, Demeterco-Berggren C, et al. (2021) Increased DKA at presentation among newly diagnosed type 1 diabetes patients with or without COVID-19: data from a multi-site surveillance registry. J Diabetes 13: 270-272. doi: 10.1111/1753-0407.13141
    [162] Vamvini M, Lioutas VA, Middelbeek RJW (2020) Characteristics and diabetes control in adults with type 1 diabetes admitted with covid-19 infection. Diabetes Care 43: e120-e122. doi: 10.2337/dc20-1540
    [163] Al Hayek AA, Robert AA, Alotaibi ZK, et al. (2020) Clinical characteristics of hospitalized and home isolated COVID-19 patients with type 1 diabetes. Diabetes Metab Syndr 14: 1841-1845. doi: 10.1016/j.dsx.2020.09.013
    [164] Unsworth R, Wallace S, Oliver NS, et al. (2020) New-onset type 1 diabetes in children during COVID-19: multicenter regional findings in the U.K.. Diabetes Care 43: e170-e171. doi: 10.2337/dc20-1551
    [165] Suwanwongse K, Shabarek N (2021) Newly diagnosed diabetes mellitus, DKA, and COVID-19: causality or coincidence? A report of three cases. J Med Virol 93: 1150-1153. doi: 10.1002/jmv.26339
    [166] Dimeglio LA, Albanese-O'neill A, Muñoz CE, et al. (2020) COVID-19 and children with diabetesdupdates, unknowns, and next steps: first, do no extrapolation. Diabetes Care 43: 2631-2634. doi: 10.2337/dci20-0044
    [167] Cameron FJ, Scratch SE, Nadebaum C, et al. (2014) Neurological consequences of diabetic ketoacidosis at initial presentation of type 1 diabetes in a prospective cohort study of children. Diabetes Care 37: 1554-1562. doi: 10.2337/dc13-1904
    [168] Kamrath C, Mönkemöller K, Biester T, et al. (2020) Ketoacidosis in children and adolescents with newly diagnosed type 1 diabetes during the COVID-19 pandemic in Germany. JAMA 324: 801-804. doi: 10.1001/jama.2020.13445
    [169] Rabbone I, Schiaffini R, Cherubini V, et al. (2020) Has COVID-19 delayed the diagnosis and worsened the presentation of type 1 diabetes in children? Diabetes Care 43: 2870-2872. doi: 10.2337/dc20-1321
    [170] Powers AC, Aronoff DM, Eckel RH (2021) COVID-19 vaccine prioritisation for type 1 and type 2 diabetes. Lancet Diabetes Endocrinol 9: 140-141. doi: 10.1016/S2213-8587(21)00017-6
    [171] Boddu SK, Aurangabadkar G, Kuchay MS (2020) New onset diabetes, type 1 diabetes and COVID-19. Diabetes Metab Syndr 14: 2211-2217. doi: 10.1016/j.dsx.2020.11.012
    [172] Kuchay MS, Reddy PK, Gagneja S, et al. (2020) Short term follow-up of patients presenting with acute onset diabetes and diabetic ketoacidosis during an episode of COVID-19. Diabetes Metab Syndr 14: 2039-2041. doi: 10.1016/j.dsx.2020.10.015
    [173] Marchand L, Pecquet M, Luyton C (2020) Type 1 diabetes onset triggered by COVID-19. Acta Diabetol 57: 1265-1266. doi: 10.1007/s00592-020-01570-0
    [174] Caruso P, Longo M, Esposito K, et al. (2020) Type 1 diabetes triggered by covid-19 pandemic: a potential outbreak? Diabetes Res Clin Pract 164: 108219. doi: 10.1016/j.diabres.2020.108219
    [175] Goyal H, Sachdeva S, Perisetti A, et al. (2021) Hyperlipasemia and potential pancreatic injury patterns in COVID-19: a marker of severity or innocent bystander? Gastroenterology 160: 946-948.e2. doi: 10.1053/j.gastro.2020.10.037
    [176] Patnaik RNK, Gogia A, Kakar A (2020) Acute pancreatic injury induced by COVID-19. IDCases 22: e00959. doi: 10.1016/j.idcr.2020.e00959
    [177] Bruno G, Fabrizio C, Santoro CR, et al. (2021) Pancreatic injury in the course of coronavirus disease 2019: a not-so-rare occurrence. J Med Virol 93: 74-75. doi: 10.1002/jmv.26134
    [178] Ghosh A, Gupta V, Misra A (2020) COVID19 induced acute pancreatitis and pancreatic necrosis in a patient with type 2 diabetes. Diabetes Metab Syndr 14: 2097-2098. doi: 10.1016/j.dsx.2020.10.008
    [179] Alharmi RAR, Fateel T, Sayed Adnan J, et al. (2021) Acute pancreatitis in a patient with COVID-19. BMJ Case Rep 14: e239656. doi: 10.1136/bcr-2020-239656
    [180] Wang F, Wang H, Fan J, et al. (2020) Pancreatic injury patterns in patients with coronavirus disease 19 pneumonia. Gastroenterology 159: 367-370. doi: 10.1053/j.gastro.2020.03.055
    [181] Pribadi RR, Simadibrata M (2021) Increased serum amylase and/or lipase in coronavirus disease 2019 (COVID-19) patients: is it really pancreatic injury? JGH Open 5: 190-192. doi: 10.1002/jgh3.12436
    [182] Coate KC, Cha J, Shrestha S, et al. (2020) SARS-CoV-2 cell entry factors ACE2 and TMPRSS2 are expressed in the microvasculature and ducts of human pancreas but are not enriched in β cells. Cell Metab 32: 1028-1040.e4. doi: 10.1016/j.cmet.2020.11.006
    [183] Kusmartseva I, Wu W, Syed F, et al. (2020) Expression of SARS-CoV-2 entry factors in the pancreas of normal organ donors and individuals with COVID-19. Cell Metab 32: 1041-1051.e6. doi: 10.1016/j.cmet.2020.11.005
    [184] Müller JA, Groß R, Conzelmann C, et al. (2021) SARS-CoV-2 infects and replicates in cells of the human endocrine and exocrine pancreas. Nat Metab 3: 149-165. doi: 10.1038/s42255-021-00347-1
  • This article has been cited by:

    1. Martin P. Arciga-Alejandre, Jorge Sanchez-Ortiz, Francisco J. Ariza-Hernandez, Gabriel Catalan-Angeles, A Multi-Stage Homotopy Perturbation Method for the Fractional Lotka-Volterra Model, 2019, 11, 2073-8994, 1330, 10.3390/sym11111330
    2. Jinglei Tian, Yongguang Yu, Hu Wang, Stability and Bifurcation of Two Kinds of Three-Dimensional Fractional Lotka-Volterra Systems, 2014, 2014, 1024-123X, 1, 10.1155/2014/695871
    3. Yukihiko Nakata, Yoshiaki Muroya, Permanence for nonautonomous Lotka–Volterra cooperative systems with delays, 2010, 11, 14681218, 528, 10.1016/j.nonrwa.2009.01.002
    4. Xiangdong Xie, Fengde Chen, Kun Yang, Yalong Xue, Global Attractivity of an Integrodifferential Model of Mutualism, 2014, 2014, 1085-3375, 1, 10.1155/2014/928726
    5. Mengqing Zhang, Qimin Zhang, Jing Tian, Xining Li, The asymptotic stability of numerical analysis for stochastic age-dependent cooperative Lotka-Volterra system, 2021, 18, 1551-0018, 1425, 10.3934/mbe.2021074
    6. YUKIHIKO NAKATA, PERMANENCE FOR THE LOTKA–VOLTERRA COOPERATIVE SYSTEM WITH SEVERAL DELAYS, 2009, 02, 1793-5245, 267, 10.1142/S1793524509000716
    7. Changyou Wang, Linrui Li, Qiuyan Zhang, Rui Li, Dynamical behaviour of a Lotka–Volterra competitive-competitive–cooperative model with feedback controls and time delays, 2019, 13, 1751-3758, 43, 10.1080/17513758.2019.1568600
    8. Ahmadjan Muhammadhaji, Zhidong Teng, Mehbuba Rehim, Dynamical Behavior for a Class of Delayed Competitive–Mutualism Systems, 2015, 23, 0971-3514, 281, 10.1007/s12591-014-0226-6
    9. Xinyuan Liao, Xianhua Tang, Shengfan Zhou, Existence of traveling wavefronts in a cooperative systems with discrete delays, 2009, 215, 00963003, 1806, 10.1016/j.amc.2009.07.032
    10. Guichen Lu, Zhengyi Lu, Yoichi Enatsu, Permanence for Lotka–Volterra systems with multiple delays, 2011, 12, 14681218, 2552, 10.1016/j.nonrwa.2011.03.004
    11. Changyou Wang, Linrui Li, Yuqian Zhou, Rui Li, On a delay ratio-dependent predator–prey system with feedback controls and shelter for the prey, 2018, 11, 1793-5245, 1850095, 10.1142/S179352451850095X
    12. Zhao Jiang, Azhar Halik, Ahmadjan Muhammadhaji, Dynamics in an n-Species Lotka–Volterra Cooperative System with Delays, 2023, 12, 2075-1680, 501, 10.3390/axioms12050501
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(4357) PDF downloads(120) Cited by(11)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog