The 2012 report by Food and Drink Administration (FDA) stated that modest rise in glycosylated hemoglobin and fasting blood glucose could result from statins therapy. This systematic review assessed and, where possible, estimated the possible association between statins use and type 2 diabetes incident. Magnitude and potential mechanism of statins-T2DM association were also of importance.
The analysis focused on review of past studies that examined statins treatment and subsequent diabetes incidents. The databases used to retrieve relevant articles included PubMed, Scopus, Science Direct, Wiley and Google Scholar. Search keywords such as “statins and incident of type 2 diabetes,” among others were applied. To be selected, a study must have been published between 2012 and 2020. Also, the articles had to be written in English.
After screening and eligibility assessment, 8 studies were included. Hazard ratios, odd ratios, and confidence intervals of each study were collected for further analysis. Odd ratios mean for 3 studies showed that each statin therapy resulted in 68% increased risk of T2DM incident (OR, 1.683; 95% CI: 1.273-2.237). The significance of effect of statins treatment on development diabetes was as well confirmed by mean hazard ratio of 49% (HR, 1.494; 95% CI: 1.19-1.686) for the remaining 5 studies.
These risks showed more prominence among older adults, normotensive males, hypertensive female, and people with low physical activity. Although no proven mechanisms were found, the suggested processes are discussed.
Keywords: Statins therapy/use, type 2 diabetes mellitus (T2DM), new-onset diabetes (NOD)
INTRODUCTION
Background
Type 2 diabetes mellitus (T2DM) was anticipated to increase from 342 million in 2011 to 534 million by 2030. The treatment of dyslipidaemias prescribes to decrease low-density lipoprotein cholesterol (LDL-C) and prevent cardiovascular disease (CVD) [1]. The treatment encourages reduction of LDL-C achieved by inhibiting 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMG-CoA R). In the United States, over 25% of adults aged at least 45 years used statins between 2005 and 2008 and 56 million may be eligible for treatment [2].
Statins indicated for secondary prevention of atherosclerotic cardiovascular disease (ASCD) and familial hypercholesterolemia [bile acid sequestrants such as cholestyramine or Colesevelam include use of PCSK9 inhibitors for resistant and familial hyperlipidaemias]. Still, there are areas of debate as in primary prevention with high LDL-C, primary prevention with low risk of CVD, chronic kidney disease, people above 75 years, and heart failure [3]. American College of Cardiology / American Heart Association (ACC/AHA) classified statins into low, moderate, and high intensity [4].
Observational and meta-analyses showed statin used in patients with concomitant risk factors for diabetes was associated with DM, but the extent and mechanism(s) is/are unclear. [5]. Study information might be deficient in changing the current practice worldview, and clinicians should screen for episode diabetes in patients on statins [6]. Therefore, extensive studies are needed to elucidate both the association between new-onset diabetes (NOD) and statin use and the underlying mechanisms [5].
Many physicians prescribe statins as recommended by ACC/AHA [7]. The adverse effects of statins are many and familiar such as muscles weakness, hepatotoxicity, and nephrotoxicity. The mechanisms of developing DM include inhibiting adipocyte maturation, expression of insulin-sensitive glucose transporter, and reduces glucose-induced insulin secretion by attenuating the glucose-induced increase in intracellular calcium concentration in vitro [8].
Summary of Previous Research
Numerous mechanisms have been suggested for statin-correlating diabetes risk, either increased insulin resistance or impaired insulin secretion. Genetic polymorphisms with reduced HMG CoA reductase function is linked with weight gain, insulin resistance and diabetes insignificant [9]. In justification for the use of statins in primary prevention, an intervention trial evaluating rosuvastatin trial (JUPITER), rosuvastatin was associated with a 27% expanded risk of NOD compared with placebo.
In contrast, the West of Scotland coronary prevention study (WOSCOPS) proposed that patients taking pravastatin confronted a 30% lower danger of diabetes compared with placebo (RR 0.7, 95% CI 0.5 to 0.99) [10]. Intensive statin therapy is more diabetogenic than moderate intensity [11]. The statins risk for NOD may differ between drugs. Therefore, conflicting data exist regarding the diabetogenic effects of statins. [12].
Recent data suggests that the use of statins may be associated with the appearance of NOD, the risk of T2DM was increased by 9% compared to placebo (OR 1.09; 95%CI 1.02–1.17) and the risk rises with increased age. Also, (statin vs control, and high-dose vs low-dose statin) without a diag nosis of T2DM at baseline. Increased chance of T2DM associated with statin use by 12% (OR 1.12; 95%CI 1.06-1.1.8) [13].
Potential clarifications incorporate remaining confounding components, for example, improved survival with statin treatment and an improved lifestyle after CV events so, the onset of diabetes-related to statin use might still be justifiable. However, independent predictors for statin related T2DM raised the degrees of benchmark fasting plasma glucose, body mass index, hypertension, and fasting triglycerides. So, there is an overlap between risk factors for developing T2DM and those same factors requiring statin usage.
A few statins have been related with expanded glycosylated hemoglobin (HbA1c) in patients accepting intensive treatment. Different statins have exhibited neutral or sound effects on glucose control in patients with and without T2DM [12]. A meta-analysis of 5 studies showed an increase in the risk of DM by 12% (OR 1.12, CI 1.04-1.22) [14]. A meta-analysis of 13 RCTs involving 91,140 participants without diabetes on statin showed prevention of five first CVD events but increased NOD by 9% [representing one additional case of diabetes per 1000 patients].
In 2012, the FDA published a safety update indicating that statins can cause a modest increase in fasting blood glucose and HbA1c [15]. Statins use have a decent wellbeing record and expanded risk of causing diabetes. [16]. Statin therapy associated DM risk and cardiovascular benefit has strengths such as historical, prospective real-life analysis, use of administrative databases to avoid recall bias, and systematic and comprehensive collection of personal data, medical history, and study outcomes.
The limitations such as a retrospective analysis, adherence assessed by dispensing information, the 5-year follow-up in our study might have underestimated and 90% of the patients prescribed a low-dose statin and about two-thirds of the patients taking statins received simvastatin; therefore, we cannot speculate whether other types of statins convey different levels of DM-induced risk [17].
Rationale and Gaps
In 28.02.2012, the Food and Drug Administration received a warning the increased risk of incident diabetes with statins, but the mechanisms clarifying the possibly higher incidence of T2DM with statin treatment have not yet been distinguished. Many clinical trials showed a possible association between atorvastatin and NOD. However, some studies have demonstrated that atorvastatin did not worsen insulin sensitivity in patients with diabetes, whereas one study suggested that patients treated with atorvastatin may be at a lower risk of developing NOD [18].
Request for Assignment Help for only $8 per page
METHODS AND DESIGN
Aim
To establish/assess the possible association between the use of statins and the development of T2DM, its magnitude and possible mechanism(s)
Objectives
To undertake a systematic review of the literature relating to the use of statins and the subsequent development of diabetes.
Research Question
Do statins cause or accelerate the new-onset T2DM? And if yes, to what extent?
Are the mechanisms understood?
Study Design
This systematic review considered only published studies.
Search Strategies and Concept
The databases search of PubMed, Science Direct, Wiley, Google Scholar, and Scopus will be undertaken and followed by the analysis of the text words contained in the titles and abstracts, and of the index terms used to describe each article. Also, the identified keywords will be used and undertaken across all included databases. The reference list of key articles will be searched for additional studies and the relevant authors might be contacted as needed for more data obtaining. The timing of published articles from 2012 to 2020 and the language is English
Eligibility Criteria for Selecting Studies for the Review
Analyzed studies were published in English and investigated incident diabetes associated with statin use. The eligibility criteria was applied to all stages of the review included the title, abstract and full-text screening. The published articles were independently assessed by two reviewers to determine the inclusion and exclusion criteria developed from research questions. PRISMA Flow Diagram, Figure 1, shows the process.
The overall number of studies picked from PubMed, Wiley, Google Scholar, Science Direct, and Scopus were 53. Further Internet searches generated 13 additional studies. All these studies were listed and after removing duplicates, a total of 61 studies remained. These were then screened. Non-randomized studies were excluded as often allocate interventions’ impact in a non-random way. For such studies, therefore, containing intervention and control groups was not enough.
Such studies were also rejected because randomized trials are possible in the case of statins intervention. Other systematic reviews and meta-analyses were also excluded as the objective was to work with primary qualitative studies. However, the systematic reviews offered valuable information that aided identification of areas with adequate evidence supporting or objecting statins as an intervention, and places where more evidence is required. Studies that availed only abstracts were also excluded as detailed information on the research processes and results could not be obtained. After screening, a total of 38 studies were excluded. 23 full-texts remained and 15 were excluded for including participants younger than 18 years.
Inclusion Criteria
Articles must be peer-reviewed, Scholarly, published after 2012 [FDA received warning statin cause T2DM]
Articles are written for academic and clinical practice, relevant to the title and research question.
Nondiabetic adults aged >18 on statins use
Persons less than 18 years
The two reviewers excluded studies that failed the quality test Critical Appraisal Skills Programme (CASP) Systematic Review Checklist
Diabetic individuals
Pregnant population
Quality Assessment/Critical Appraisal
Each reviewer will independently check each selected article to determine the quality [rigour of methodology] and minimize bias. Critical Appraisal Skills Program (CASP) Systematic Review Checklist. Two different tools were applied to appraise the studies for review. The first checklist was the Cochrane risk-of-bias tool. The tool is critical for evaluation of quality and bias risk when dealing with randomized clinical studies. A copy of the checklist is attached in the Appendices section (Appendix I). This systematic review had a specific focus on randomized controlled trials (RCTs).
As explained by Harrison et al. [19], RCTs contain the highest and most reliable evidence to apply when analyzing the impact of interventions such as drug treatments. For a study to be included, therefore, it had to include two groups (where one receives the intervention and the second one is a control group). When determining the methodological quality of each of the RCTs, risk of bias was the key issue. In this case, ‘biases’ referred to errors in results or observable deviations from what is considered to be true.
Cochrane’s bias assessment tool enabled evaluation of biases during detection, selection, performance, and reporting, among others. The assessed articles were rated to have either lo, high or unclear bias risk. While investigating performance bias, studies that informed participants if they would be on the intervention or control side were excluded. Such studies had high bias risk. The checklist on other bias was crucial for determining if the RCTs included participants targeted by this systematic review (that is, persons who are 18 years or older).
The second tool is Joanna Briggs Institute (JBI) checklist for critical appraisal. Use of this tool is vital for synthesizing and interpreting the findings of a study. The JBI tool enabled evaluation of different forms of biases experienced in quantitative studies. The checklist is presented as Appendix II.
Data Collection /Extraction
Author consensus determined study inclusion in this review, focusing on observational studies, randomized control trials and meta-analyses. Data were extracted by the author in the data extraction form included the review question, citations, aims and fulfils the requirements for systematic review. The corresponding authors can be contacted in case of missing or/and ambiguous data.
Data Analysis and Synthesis
The studies were grouped; the extracted results compared the design, methods, participants, sample size, incidence, relative risk, confidence interval, and effect size. Numerical, tabular, and graphic presentation were addressed.
Ethical Issues
It is data in the public domain, so no ethics approval required
Anticipated Outcomes and Impact
The 2 points will add a positive effect:
To inform the physicians to use the statins as recommended by major clinical guidelines.
Balance and weighing the pros and cons of statin
RESULTS
8 studies met all the terms of the inclusion criteria. Table 1 below presents the details on the nation where each study took place, the purpose of the investigation (aim), the features of the studied population, and the research design applied to meet the objectives of each study.
Table 1: Location, Aim, Population Features, and Study Design
Citation | Country of Study | Aim of Study | Study Population Characteristics | Study Design |
1. Lee, Sung, Cho, Kim, and Chang (2018) | Republic of Korea | To determine if risk of diabetes mellitus increases with satin use, and further examine the how gender and hypertension influence the association between satin and DM. | Study concentrated on the health-screening cohort all over the nation. Selected individuals had a minimum cholesterol level of eve mg/dL. Also, they had to be free from pre-diagnosed diabetes, cancer, and cardiovascular disease. They must have undergone multiple health checkups. | Retrieval of demographic and eligibility details of participants from National Health Insurance System. The information comprised records of medical treatment, drug subscriptions, smoking history, alcohol consumption, and prescriptions to better blockers or hydro choro thiazide. |
2. Li, Lin, Zhao, Xu, Cheng, Shen and Zhan (2018) | China | To investigate the link between statin therapy and incidences of new-onset DM in individuals suffering from hypertension. | Patients with or without type 2 diabetes. The selected age range was 30 to 90 years. | The non-user category consisted of patients that with no experience of statin throughout the study. On the other hand, statin users received statin therapy for a minimum duration of 90 days at the time of the study (Jan 2011 to August 2016). The index date was date of first prescription to statin. Study endpoint would be marked by Type 2 DM diagnosis. Also the main endpoint was set after 180 days as researchers assumed that NOD before this period could not be linked to the treatment. |
3. Kim, Kim, and Park et al. (2019) | South Korea | To employ population-based data in the analysis of the effect of duration and recent application of statin on the possibility of NOD. | Participants were among the 1 million individuals that subscribed to medical insurance in the country of study. | The researchers studied prescription listing for over a duration of three years prior to the desired outcomes (2009 to 2013). The analysis involved following of subscribers between January 2012 and December 2013. Five groupings were applied for statin administration: not users; users within less than 6 months; use over 6 months but below 1 year; use within a year but less than 2 years; and users in 2 years but not yet 3 years. |
4. Ko, Jo, Kim, Kang, Cho, Jo, Park, Yun, Lee and Park (2019) | Korea | To evaluate the relationship existing between statin therapy application and NODM based on intensity, duration, and cumulative statins dose. | Included adults with reported hypercholesterolemia in the period Jan 2005 to December 2012. Also included were adults in their 40s and older without atherosclerotic cardiovascular disease, and overall cholesterol level exceeding 240mg/dL. | The investigation’s drug exposure of concern was the ever versus never users of statin-based therapy. Ever usage was marked by a minimum of two prescriptions within 6 months. Never usage described all patients that did not fit the ever use criteria. The patients with statin therapy use were further subdivided based on duration of utilization, cumulative dosage, and intensity status. The duration ra nged between less than a year and more than 2 years. The intensity defined that patients receiving atorvastatin dose comprising 40-80 mg/d (or 20-40mg/d) were on high-intensity dosage. |
5. Corrao, Compagnoni, Rea, Merlino, Catapano, and Mancia (2017) | Italy | To inform on the level of impact of type 2 DM (induced by statins) on risks of macrovascular complications as opposed to diabetes which does not result from statin therapy. | All the selected beneficiaries of National Health Service (NHS) belonged to Lombardy. The preferred age range was 40 to 80 years. To be included, an individual had to have a minimum of statins prescription in 2003 to 2005. | There were four patient classes including: (i) those prescribed 1 or more statins therapy, 36 months before the index date for step-1; (ii) patients with a minimum of a single antidiabetic drug, or hospitalization following diabetes diagnosis 36 months to the index date; (iii) patients hospitalized due to cardiovascular disease diagnosis, or those that received related drugs (nitrate, etc.) 36 months before the index date for step-1; and (iv) patients that failed to renew first statins prescription as well as others that did not attain 12 months follow-up. For minimization of false positive diabetic incidents, 3 prescriptions of antidiabetic drugs were necessary. This was required before ascertainment of the outcomes of step-1. |
6. Yoon, Sheen, Lee, Choi, Park, Rae , and Lim (2016) | Korea | To examine if using statins increases NODM risk among Koreans. | Study population comprised both statin and non-statin users age 18 years and above. | To be considered statin-exposed, one must have used the therapy consecutives for over 90 days. Also a repeated drug prescription should have occurred in 30 days. For the non-exposed, participants must have been followed for over 90 days and were not found to have any satin exposures. The exposed patients were sub-categorized into 6 classes defined by the type of statin used. Statin exposure observations started on the 91st day. The observation outcomes did not include the ones for patients already diagnosed with diabetes. |
7. Anyanwagu, Mamza, Donnelly, and Idris (2017) | United Kingdom | To investigate how statins affect glycemic reaction and mortality results after type 2 diabetes patients commence insulin therapy. | Included individuals diagnosed with type 2 diabetes who are more than 18 years old. Also patients that began insulin therapy between Dec 2006 and May 2014 were considered. | Follow-ups were conducted on every patriate that began insulin therapy from when the statin treatment began. Outcomes were compared against those of others that only started statins experiencing the secondary outcomes. |
8. Ahmadizar, Ochoa-Rosales, Glisic M, Franco, Muka, and Stricker (2019) | Netherlands | To examine the connection within statin therapy and glycemic factors and type 2 diabetes incidents | Were residents of Ommoord aged 45 years or older | The study employed 2 phases of analysis. The first was a cross-sectional assessment of link between statins and glycemic properties. The second stage comprised of longitudinal follow-ups to evaluate determine how statins associate with type 2 diabetes. |
Table 2 identifies the research setting, the selected sample size, techniques applied during sampling, and the places from where the researchers collected the data to accomplish the study requirements.
Table 2: Setting, Size of Sample, Sampling Approaches, and Sources of Data
Citation | Study Setting | Sample Size | Sample Technique | Data Source |
1. Lee et al. (2018) | A significant rise in prevalence of T2DM was noted in South Korea between 1970 and 2000. Also, the country was ranked among the countries with enormous numbers of adult diabetes (for persons aged 20 to 79 years). This made the South Korea one of the 30 members of Organization for Economic Cooperation and Development. The country is among those with highest population growth rates and is, thus, projected to record the largest numbers of T2DM come 2030. It was also noted that cardiovascular disease prevalence had more than doubled that of diabetes. The revelations indicate a need to plan national health interventions and identify the potential factors for type 2 DM incidence. | n= 40,164 patients (17,798 used statin and 22,366 did not use statin.) | Random sampling of people that began biennial NHIS health checkups around 2002 to 2003. | The National Health Insurance System (NHIS) database. Data reported in 2016 was used. It represented 10% of the overall target population. |
2. Li et al. (2018 | Yinzhou district in Ningbo city, China. The Health and Family Planning Commission of Yinzhou declared its intention to adopt primary care framework used in the United Kingdom, in 2005. Therefore, apart from being the largest district, Yinzhou has extensive data on diabetes, cancers, and hypertension follow-ups. | n= 67,993
(21,551 were statin users while 46,442 were non-users) | Random sampling of hospital, public health and community health data contained in the Health Information System within Yinzhou. | Yinzhou regional healthcare database. Data recorded between Jan 2010 and August 2016. |
3. Kim et al. (2019) | The nationwide case control study conducted with approval of Korea University’s review board. | 6417 statin users and 32,085 control group | Stratified random sampling of data from the database. | South Korean National Health Insurance National Sample Cohort database. The database includes a randomized sample consisting of 2% of the overall country population. |
4. Ko et al. (2019) | Examined combined data from the national health insurance and datasets based on statements of the administration. The study targeted the Korean National Insurance Services because the organization had the obligation to collect and manage all the relevant information. As such, the possibility of data loss would be minimal. Researchers were, thus, certain that their study objectives would be obtained through the systematic linkage codes used in in- and out-patient points along with the records showing the dispensed drug prescriptions. | n = 2,162,119 (638,625 eligible for statin therapy and 1,523,494 never tried statin) | Purposive sampling of adults with hypercholesterolemia, and random sampling of participants that fit this group. | National Health Insurance Services for Korean population |
5. Carrao et al. (2017) | The study targeted Lombardy region, Italy. The region consists of 16% of Italy’s population covered by NHS | 4,391 eligible diabetic division and 77,893 other group members | Random selection of diabetics and others without diabetes symptoms. | NHS healthcare databases of Lombardy |
6. Yoon et al. (2016) | Ajou University Hospital, a tertiary teaching institution found in Korea. | 8265 participants already exposed to statins and 33,060 patients with no statins exposure in the study period (Jan 1999 to Aug 2013). | Random sampling of patients whose demographic, diagnosis, drug subscription and laboratory test outcomes were recorded in the database. | Clinical research database belonging to Ajou University Hospital |
7. Anyangwagu et al. (2017) | The retrospective group investigation focused on the UK primary care setting. | Consisted of 10,682 previous statin users and 2043 that never used statins | Random selection of samples from the longitudinal records of about 587 general practices. | The Health Improvement Network (THIN) database within an extensive electronic primary care in the UK. |
8. Ahmadizar et al. (2019) | Study targeted Ommoord district in the Netherlands | 9535 participants for analysis of statins and glycemic factors, and 8567 patients for statin and type 2 diabetes development | Random selection of samples from the dispensed data. | Rotterdam study reports on visits of 1997 to 1999, 2000-2001 and 2006 to 2008. |
Table 3 indicates the various factors measured in each of the 8 selected studies. It also illustrates the methods employed during data analysis, the confounder variables, and the main observations made for each study.
Table 3: Measures, Methods of Analysis, Confounder Variables, and Primary Observations
Citation | Measures | Analysis | Confounder Variables | Key Observation |
1. Lee et al. (2018) | Fasting glucose, total cholesterol, systolic and diastolic blood pressure, body mass index (BKMI), height, body weight, and waist circumference. | Percentages, means, Chi-square tests, and standard deviations applied on normally distributed variables. Median for non-normally distributed. Cox regression for rate of new-onset DM development and Hazard raters to calculate influence of statin on DM development. | Female gender, hypertensive women, and normotensive men | Statin use was a risk factor for normotensive patients as well as hypertensive women. |
2. Li et al. (2018) | Age, sex, BMI, comorbidities, features of lifestyles and baseline antihypertensive drugs in use. | Employed multivariate Cox model alongside propensity approaches to identify potential confounders. Cox proportional hazard framework to relate statin initiators with non-users. Propensity scoring of statin use were determined through logistic regression. The analysis also applied Cox regression. Mean, standard deviation and median for BMI and age. Also, student’s t test for variations in BMI and age. | Use of antihypertensive drugs, older adults, overweight, and having dyslipidemia | Found a significant relationship between use of statin and rising risk of new-onset DM. |
3. Kim et al. (2018) | The inspection utilized self-questionnaire to measure drinking, smoking, and exercise status of participants. | t tests for continuous variables and Chi-square for categorical data. Also, conditional logistics model for regression aided the estimation of relative usefulness of statin therapy. | Gender, smoking, alcohol consumption, high total cholesterol, poor exercising routine, the first six months of statin therapy and hypertension. | Recent and short term use of statin was linked to a rise in risk of NOD. Cumulative duration in use of statin and non-recent application of statin therapy had no significant association with NOD. |
4. Ko et al. (2019) | Impaired fasting glucose, BMI of 25kg/m2 and above, effect of different types of statin (atorvastatin, rosuvastatin, simvastatin, pravastatin, etc.), clinical risk of DM development | Propensity score applied to measure differences in baseline features of statins users and non-user groups. Student t test was applied on continuous variables. Study also employed Cox regression framework to compare outcomes of risks. For estimation and comparison of proper approaches, Kaplan-Meier method was used. | Older age, gender, type of statin. | Findings revealed time and dose-dependent connections in statin use and indicated that the association increased the risk of NODM. |
5. Corrao et al. (2017) | Calculation of period of prescription, identification of statins type, determining existence of antihypertensive agents in the past, test for fibrates, test for chronic obstructive lung disease, and comorbidities. | Involved a two-stage data evaluation. First stage included search for replication of findings on factors that raise chances of diabetes development. Also used Cox proportional model to determine ratio of hazard. In the next step, the chances (greater/lesser) of diabetes development were derived. Two ancillary assessments to find out the cause of macro vascular complications. | High adherence to statins, male gender, older adults, combination of atorvastatin with antihypertensive, anti-inflammatory, respiratory and antithrombotic agents. | With very high level of previous statins adherence, type 2 diabetes could not be associated with macro vascular risk. This implied that Type 2 DM was induced by statins. |
6. Yoon et al. (2016) | Baseline glucose levels, gender, hypertension, age, comorbidities, concomitant drugs | For continuous variables, study employed mean, standard deviation and percentages for categorical variables. Chi-square and t tests enabled comparison of exposure levels across groups. NODM risk was tested by Kaplan-Meier while Cox proportional hazard regression model tested effect of statin exposures. | Older age, male gender, greater level of baseline glucose, thiazide exposure, and hypertension | An increased risk for new-onset DM for statin users. Patients exposed to atorvastatin showed a higher NODM risk than those without statin use. However, no significantly different outcomes were found with statins type comparisons. |
7. Anyanwagu et al. (2017) | Glycemic control, body weight changes, systolic and diastolic blood pressure (for 6 months, 1 year, 2 years, and 3 years) since starting insulin treatment, age, gender, socioeconomic situation, smoking and consumption of alcohol, body weight, height, and biochemical parameters. | Mean, standard deviation, and baseline features to compute summary variables. Pearson’s Chi-square and t test enabled determining of differences in treatment groups. Cox proportional hazard model was employed for related ratios and Kaplan-Meier measured survival elements in treatment groups. Cox regression model checked for breach of assumptions on proportional hazard. | Concurrent statins use | After the commencement of insulin therapy, concurrent application of statins reduced glycemic control in the short run. It also minimized the chances of the increasingly fatal incidents of cardiovascular disease. |
8. Ahmadizar et al. (2019) | Serum glucose, insulin concentrations, health status, BMI, hypertension, height, weight, total cholesterol, diastolic/systolic blood pressure, triglycerides, and lipoproteins-cholesterol. | Chi-square statistics for relating statin use and t tests for continuous variables with normal distribution. Also univariate and multivariate linear models for regression analysis of statins and glycemic relations. Correlation coefficients helped with identification of level of connection between insulin concentrations and serum fasting glucose | Cumulative exposure | There is higher hyperglycemia risk among statin users. This could cause resistance to insulin and in the long run lead to the development of type 2 diabetes. |
Further Results
After determining the effect of statins on T2DM, as indicated in the Key Observations of every study, this systematic review took further steps to determine the extent of the effect statins have on T2DM. For this objective, confidence intervals and comparison percentages (for statins users and non-users) were collected. The intervals are presented in the table below.
Table 4: Calculation of the Odd Ratios and Confidence Intervals
From Table 4, each statin therapy is associated
with 68% increase in risk of T2DM incident (OR 1.683; 95% CI: 1.27-2.237). Data on confidence intervals collected and presented above were used to construct the histogram in Figure 2.
The risk of T2DM with statin use ranges between 1.21 and 2.45. Median value was 1.21-2.44.
Figure 2: Confidence Intervals from included Studies
The remaining 5 studies had Hazard ratios and confidence intervals presented in Table 5.
Table 5: Hazard Ratios and Confidence Intervals Calculation
The findings above show that statins use increases chances of T2DM incident by 49% (HR 1.494; 95% CI: 1.19-1.686).
Figure 3: Confidence Intervals for Studies with Hazard Ratios
Request for Assignment Help for only $8 per page
DISCUSSION
In order to respond appropriately to the research question, this section covers three main areas: (i) the possible association between statins use and incidents of T2DM; (ii) Magnitude of the association; (iii) Possible mechanisms.
Association between Statins Therapy and T2DM Development
Findings of various studies were partly consistent with each other. All the included studies reported a significant association between statin and the new-onset diabetes mellitus even though the conditions that contributed to the risk of NODM differed. For instance, Lee et al. found that statin therapy raised chances of NODM but risk appeared more common among normotensive participants and hypertensive females.
Lee et al. [20] described T2DM as a sophisticated disease with several risk factors. With the involvement of metabolic syndrome in development of this disease, tracking the actual impact of statins might turn out challenging. To avoid misleading conclusions, they recommended that observers check for impaired fasting glucose rates, and it could be useful to test for metabolic syndrome prior to statin therapy. Some of the associated risk factors included levels of fasting serum glucose and triglyceride, BMI, and hypertension. These risk factors were proven in RCTs that involved atorvastatin.
With hypertension, however, several clustering elements beyond statins contribute to the development of DM. In fact, statin therapy had no connection to DM among hypertensive men. Although these findings shared points with previous ones that attributed popularity of T2DM to female and not male gender, Lee et al. could not state with confidence if the gender differences was risk or if other factors influenced how men and women used statins. However, the discovery necessitates separation of treatment and monitoring approaches across gender.
On the other hand, Li et al. [21] found an association between statin therapy and higher possibility of NOD among hypertensive patients, compared to others that did not use the drug. The researchers compared their findings to the outcomes of previous studies and concluded that the effect of statin in their study was moderate. Li and colleagues explained that the higher levels of NOD in their study could have been caused by the fact that they targeted patients with hypertension, who could be at greater risk for DM compared to the general population.
Although previous analyses associated NOD with statin use more prevalent among women, Li et al. found that the risk of NOD was consistent for both genders across age groups. The relationship was more significant among older patients. Some of the more stable findings in the study by Li and colleagues included that T2DM incidents rose with age for statin non-users but the risk stayed high throughout for the users. Li et al. also found that impact of statin on NOD incidence varied with the specific type.
In particular, atorvastatin and simvastatin showed increased risk of NOD based on the results of multivariate Cox models analysis. They further revealed that NOD risk grew with rise in intensity of statins use. With statin therapy, also, cumulative number of days and non-current statin use had no contribution to NOD risk [22]. In this study, NOD risk only grew with statin among individuals who engaged in therapy within and below 6 months. Ko et al. reported a time- and dose-based interrelation of use of statin and rising risk of incident DM. Their findings, however, contradicted the ones by Lee et al in that gender did not play a role in raising the chances of DM incident. In their study also, age [21] and existence of risk factors [22] had no relevance.
Additionally, the type of statin treatment did not cause variations in NOD risk. A factor that increases the relevance of the research by Ko and colleagues is that they witnessed the complete reimbursement of statin interventions for individuals with hypercholesterolemia [23]. They were present at the primary prevention point, and took note of the observable changes. The information provided by the study can thus be of clinical relevance. In the inquiry by Corrao et al. [24] patients with near complete statins adherence (or above 90%) NOD risk was close to the non-diabetic persons but the rate was double the one for persons without diabetes for the patients with extremely low adherence (or adherence lower than 10%).
The progressive decline in DM risks which happened with increased adherence to statins revealed a higher protective influence of the drug in diabetic patients using the therapy. The findings imply that continuous application of the statins therapy may attenuate and even counterbalance the drugs’ diabetogenic effect so that the shortcomings of their administration are controlled. Yoon et al.’s study yielded relatively higher hazard ratio (HR) than previous studies.
The authors attributed the higher values to the possibility that their investigation brought together patients with both dyslipidemia and comorbid cardiovascular disease. The chances of this are even higher as the inquiry concentrated on patients attended to within a tertiary hospital [25]. Yoon and colleagues explained that the chances of applying their findings to a larger patient cohorts were minimal but their results were of clinical significance because cardiovascular disease patients require statin therapy.
Anyanwagu et al. reported that type 2 diabetes patients initiated with insulin became more complex, showed longer disease duration, and increased cardiovascular disease risk. These findings were crucial as they revealed the implications of statin treatment on cardiovascular and metabolic outcomes, after 5 years of insulin use [26]. Further, Anynwagu and others found a more impressive Hb1c lowering effect in non-users of statin compared to the users.
The results were obtained after the two groups began insulin therapy. Although no weight gain differences were observed, systolic blood pressure appeared lower for the non-statin users. The study worked with older participants who lived longer with diabetes and used insulin treatment more than the included non-users.
Ahmadizar et al. indicated that elevation of a series of glycemic features were a major observation of statin association at baseline [27]. Even with adjustments in age, exercising routine, and educational attainment, the observed influence of statins on DM development persisted.
Magnitude of Statins Use and T2DM Incidents
Each of the chosen articles contained an effect size reported by the researchers. In the study by Lee et al. the rate of T2DM occurrence among statins users was higher (7.6%) compared to the non-users, which was 5.7% [20]. On the other hand, Li et al. indicated that the risk of developing T2DM increased by 54% for statins users [21]. Ko et al. compared the effect and reported that the impact between users and non-users of statins was 13.4 and 6.9 per 1000 person-years, respectively [22].
Similarly, Yoon et al. found the effect rate to be 6.0 and 3.2 per 1000 person-years for statins users and non-users, respectively [23]. The article by Anyanwagu et al. showed the difference in T2DM incidents to be 20.7 for non-users and 30.9 for users, per 1000 person-years [24]. Kin et al. found that only use of statins in the short run contributed to the emergence of new-onset DM. The effect was 1.48 for the short-term users and none for non-users [25].
Corrao et al. reported that the risk of T2DM was 24% among patients with low statins adherence, 72% for the intermediate, and 95% for those with high adherence to statins [26]. Finally, Ahmadizar et al. found that cumulative use of statins increased the chances of T2DM emergence by 38% [27].
Possible Mechanisms in Statins Use and Development of T2DM
Lee et al. [20] noted that many studies already found a significant association between statins therapy and DM incidence, the relevant mechanisms were not yet properly characterized. They stated that many potential mechanisms were in existence. One such suggested mechanisms indicated that statins may directly cause a decline in the amount of synthesized insulin or impair its secretion.
The therapy could also heighten insulin resistance or even transform insulin signaling within peripheral tissues. There is one other proposed mechanism which was not applicable in Lee et al.’s investigation. It suggested that the interaction between statins and lower cholesterol levels could induce DM. In this mechanism, the effect should not be attributed to the drug alone. For Lee et al. inquiry, however, adjustments in cholesterol levels did not interfere with statins ability. In other words, statins influenced DM development by itself.
Another mechanism states that the expected age gradient would be moderated if younger individuals with DM development risk were correctly selected for statin interventions [21]. In the analysis, patients with a BMI below 24 were identified to face higher diabetes risks compared to individuals with 28 and beyond. Although the findings were similar to those of research on postmenopausal females, Li et al. recommended further investigations to examine the influence of weight on statin-related new-onset DM.
Kim et al. noted the existence of biologic mechanisms for induction of diabetes by statin. One such mechanism postulated that the inhibiting impact of statin concentrates on 3-hydroxy-methylglutaryl-CoA reductase. The genetic variants within this element and statin intervention were found to cause increased body weight and higher possibility of T2DM. They also indicated that chances of NOD incidents increase with reduction in levels of low-density lipoprotein (LDL-C).
Statin therapy may as well interfere with functioning of beta cell. On its own, statin therapy reportedly has the ability to destroy ß cell. Additionally, statin use minimizes expression of glucose transporter, thereby interfering with uptake of glucose to the peripheral cells. Statin treatment could also cause accumulation of cells by inhibiting the differentiation of adipocyte. This prevents the secretion of hormones that are sensitive and resistant to insulin.
On the other hand, lower indicators of diabetic condition included predisposition of HbA1c, familial hypercholesterolemia, fasting blood glucose (FBG), higher levels of blood LDL-C, dyslipidemia, and homeostatic model assessment of insulin resistance (HOMA-IR) [22]. Ko et al. indicated that excessive DM risk is only prevalent among patients with the primary risk factors for the disease. They stated that a standard statin dose would contribute a 10% proportional rise in reported DM. Yoon et al. also noted that atorvastatin had a considerable effect on NODM development. Atorvastatin is a lipophilic statin type, hence contains extra diabetogenic potential that manifest through the many harmful side-effects.
Also, statins in this category lack glucose tolerance. They have no difficulty penetrating the cell membrane, and this capability places their extrahepic impact above the hydrophilic statins. Simvastatin also belongs to this group. Together with atorvastatin, the drugs inhibit L-type Ca2+ pathways and interfere with exocytosis, within pancreatic beta cell [25]. The outcomes include reduction in insulin secretion, decrease in glucose transporter 4 expression, and increased insulin sensitivity.
Anyanwagu et al. focused more on the ways statins worsen dysglycaemia. One mechanism, thus, involved the direct/indirect effect of statins on calcium pathways that led to decrease of insulin secretion as a result of pancreatic better cell impairment. Another way is impairment of glucose transporter 4, which result in a drop in uptake of insulin-mediated glucose. This happens mostly in the skeletal muscles.
Statins treatment could also deplete coenzyme Q10 so that phosphorylation activities are inhibited in the lower insulin receptor [26]. This may also prevent adipocyte differentiation. With these occurrences, statins increases muscle resistance of insulin. Ahmadizar et al. found various mechanisms through which statins act on individuals free from diabetes. One is a decline in uptake of insulin-mediated cellular glucose, which results in intolerance for glucose.
Secondly, there occurs a decline in isoprenoids synthesis that lowers regulation of glucose transporter 4. The final outcome is hyperinsulinemia, and hyperglycemia. The third mechanism happens with reduction of farnesyl pyrophosphate, coenzyme Q10, and dolichol [27]. These cause alteration in secretion and resistance of insulin.
CONCLUSION
The systematic review found a significant level of risks of statin therapy on NOD incident. The analysis also revealed that statin therapy increased the possibility of NOD incidents among users than non-users with hypertension [20] [21] or those that were normotensive [20]. This situation persisted even with differences in characteristics of patients, statin types, and levels of dosage. Contrarily, the risk was quite minimal in old and obese patients, with increased risk of cardiovascular disease [21]. Generally, statins benefits exceed the related risks more so for individuals that face higher risk of or those already diagnosed with cardiovascular disease.
It is, therefore, necessary that proper guidelines be followed by clinicians when prescribing statins therapy. However, the advantages and disadvantages of statins use must be considered for patients below 50 years [21]. On the other hand, Kim et al. raised concern over the current statin dosage, explaining that it has high potential of increasing NODM risk. The authors argued that short-term diabetes risk caused by statin treatment showed more prevalence among patients exposed to risk factors for diabetes [22]. That is, effects of statin intervention like induction of insulin resistance, decline in expression of glucose transportation, and functionalities of beta cell, could heighten the chances of developing diabetes.
A common strength shared across the eight studies is that they collected data from reputable sources and included the most relevant factors in the study. In all the studies details such as gender, age, BMI, waist circumference, weight and height were taken. Others also included alcohol consumption, exercising routine, LDL-C, and hypertension [20] [22]. The studies also examined the connection between NODM risk and statin intervention using large-scale databases.
Policy Recommendations
More research is required in countries with high and growing numbers of obesity, cardiovascular disease and diabetes. Findings of such studies will inform national health planning strategies through identification of the actual factors that lead to the hike in T2DM incidents. Further research in this area could also promote identification of patients facing increased risk of T2DM, and provide guidance for treatment of such individuals in clinical contexts.
For instance, Lee et al. [20] argued that the modifications included in the blueprint for dyslipidemia management could encourage the adoption of statins therapy across the globe. This study also indicated that women and normotensive males faced higher risk for DM. They suggested serious monitoring and follow-ups for patients in these categories. It is necessary to check the levels of HbA1c and FBG prior to the commencement of statin treatment. This might be useful when handling prediabetes individuals that have high risk of diabetes [22].
Diabetes screening is as well of importance to patients requiring statin treatment. An amazing discovery was brought forth by Corrao and colleagues. The researchers found that statins were crucial in cardiovascular treatment and showed impressive results in patients with diabetes and those that only had cardiovascular disease [24]. This means that practitioners should not withhold statins treatment from patients suffering from both diabetes and cardiovascular disease, as no risks are involved.
Yoon and others discussed that the benefits of statins in cardiovascular disease treatment rest on the drug’s therapeutic efficacy on hyperlipidemia and influence on pleiotropic elements. While supporting the effectiveness of statins in atherosclerotic cardiovascular prevention due to proven effectiveness and safety, Yoon et al. suggested that NODM risks associated with statins use should be labelled on the drug package or container [25]. This step will assure users that balanced insight has been considered for the treatment. The researchers indicated that investigators should check for glycemic impairment when investigating incident diabetes, in an experiment where impaired glucose metabolism and sensitivity of insulin are involved [27].
Implications for Further Research/Reviews
The reviewed studies also had limitations which could interfere with the clarity and reliability of the conclusions derived in this systematic review. For instance, the study by Lee et al. depended on observational data based on South Korean population. Given that the study subjects came from East Asia, use of the recommended clinical practices requires caution as the outcome may not be consistent in patients belonging to different ethnicities.
The use of observational data lacks control over confounder variables. This means there are higher chances of finding false links in drug and results. Li et al. [21] also employed observational data and argued that their definition of bias could promote bias. Besides, they utilized inpatient and outpatient diagnosis to select subjects suffering from diabetes mellitus, hypertension, and other comorbidities. The limitation in this results from the possibility that the records may not have captures the true diabetic patients.
This problem also applies to observational studies relying on databases for data. Li et al. also reported an encounter with confounding factors beyond their control. The study also failed to acquire adequate information on hypertension and dyslipidemia levels. This denied the research the opportunity to work with these crucial potential T2DM risk elements. Again, Yinzhou database did not include dietary systems which influence T2DM incidents [21]. Lastly, with the reliance on records within the database, it was impossible to tell if the indicated patients really used the medications prescribed by the doctors.
The NHIS database utilized by Kim et al. also did not include data on Hb1c. As a result, it was not possible to prove the use of definitive criteria for diagnosis of diabetes [22]. In the same study, the specific statin therapy and the applicable dosage were not identified. Although Ko et al. did not find any effects related to specific type of statin, they cautioned practitioners against selecting a strain of the drug for populations known to have high risk of DM development [23]. The reason for this is that their study findings may not be conclusive as they had access to limited data on the different risks of DM.
Although substantial adherence to statins had impressive outcomes in the study by Corrao et al., the researchers stated that other anti-diabetic and anti-hypertensive agents were involved. These great therapeutic package must have influenced the results of step-1 and 2 follow-ups. Again, the findings on beneficial effect of statins on cardiovascular disease and related drugs may not be factual as the hospitalization data applied in the analysis did not reveal values for fasting glucose, LDl-C, blood pressure, and lipids.
Also, the follow-up duration did not attain the required level for the evaluation of impact of novel statin treatment [24]. Therefore, there is a need to consider the diabetes risk factors such as unhealthy diet and BMI (which could not be found in the database chosen by the researchers).