Statin Prescription Patterns and Associated Factors Among Patients with Type 2 Diabetes Mellitus Attending Diabetic Clinic at Muhimbili National Hospital, Dar es Salaam, Tanzania

Introduction

Cardiovascular disease (CVD), which includes coronary artery disease (CAD), cerebrovascular accident, and peripheral arterial disease (PAD), is the leading cause of morbidity and mortality in patients with type 2 diabetes mellitus (T2DM).¹ Patients with type 2 diabetes have an increased prevalence of lipid abnormalities, contributing to their high risk of CVD.^2,3 Multiple clinical trials have demonstrated the beneficial effects of statins therapy on CVD outcomes in patients with and without coronary heart disease (CHD).^4,5 Subgroup analyses of patients with diabetes in larger trials^6–9 and trials in patients with diabetes^10,11 showed significant primary and secondary prevention of CVD events and CAD deaths. Meta-analyses of 14 randomized trials of statin therapy, including data from over 18,000 patients with diabetes (mean follow-up 4.3 years) demonstrated a 9% proportional reduction in all-cause mortality and 13% reduction in vascular mortality for each mmol/L (39 mg/dL) reduction in LDL cholesterol.⁴

Accordingly, statins are the drugs of choice for LDL cholesterol lowering and cardio-protection. Statin prescription patterns in patients with diabetes are categorized as moderate or high intensity dosage. Moderate intensity dosage includes: atorvastatin 10–20 mg, rosuvastatin 5–10 mg, simvastatin 20–40 mg, pravastatin 40–80 mg, lovastatin 40 mg, fluvastatin 40 mg twice a day/80 mg once a day and pitavastatin 2–4 mg whereas high intensity dosage includes: atorvastatin 40–80 mg and rosuvastatin 20–40 mg.¹² High intensity statin therapy achieves approximately a 50% reduction in LDL cholesterol and moderate intensity statin regimens achieve 30–50% reductions in LDL cholesterol. Low-dose statin therapy is generally not recommended in patients with diabetes but is sometimes the only dose of statin that a patient can tolerate.¹³ For patients who do not tolerate the intended intensity of statin, the maximally tolerated statin dose should be used.¹² As in those without diabetes, absolute reductions in atherosclerotic cardiovascular disease (ASCVD) outcomes (coronary arterial disease death and non-fatal myocardial infarction) are greatest in people with high baseline ASCVD risk (known ASCVD and/or very high LDL cholesterol levels), but the overall benefits of statin therapy in people with diabetes at moderate or even low risk for ASCVD are convincing.^14,15

The American Diabetes Association (ADA) standards of medical care in diabetes of 2021¹² recommends that: For patients of all ages with diabetes and ASCVD high-intensity statin therapy should be added to lifestyle therapy (Grade A recommendation); for patients with diabetes aged 40–75 years (grade A recommendation); for patients aged >75 years without ASCVD, use of moderate-intensity statin in addition to lifestyle therapy is recommended (grade B recommendation). For patients with diabetes aged <40 years with additional ASCVD risk factors, the patient and provider should consider using moderate-intensity statin in addition to lifestyle therapy (Grade C recommendation); In clinical practice, providers may need to adjust the intensity of statin therapy based on individual patient response to medication (for example; side effects, tolerability, low-density lipoprotein [LDL] cholesterol levels, or percent LDL reduction on statin therapy); for patients who do not tolerate the intended intensity of statin, the maximally tolerated statin dose should be used (Grade E recommendation); for patients with diabetes aged 40 years and above with multiple traditional cardiovascular risk factors without obvious ASCVD high intensity statins to be considered (Grade C recommendation).

Despite the fact that the role of statins in both primary and secondary prevention of cardiovascular disease is established, several real-world studies have shown a distance between guideline recommendations and use of statin treatment.^16–18

Tanzania, like other developing countries, is facing a high burden of cardiovascular diseases (CVDs). The country is experiencing rapid growth of modifiable and intermediate-risk factors that accelerate CVD mortality and morbidity rates. In rural and urban settings, diabetes is among the cardiovascular risk factors which contribute to mortality and morbidity.^19,20 High cholesterol is one of the most important and modifiable risk factors for the development of cardiovascular disease.²¹ Appropriate prescribing of cholesterol-lowering medications, therefore, represents one of the most important and accessible interventions for combating the burden. Obtaining baseline data on patterns of statin prescriptions, prevalence, and adherence to dosages of statins by benefit groups, sociodemographic and clinical determinants of statins prescription is critical for future intervention and prospective studies. The aim of this study was to determine statin prescription patterns and associated factors among type 2 diabetes patients attending a diabetes clinic at Muhimbili National Hospital, a tertiary hospital in Dar es Salaam, Tanzania.

Materials and Methods

Study Design, Area and Setting

A cross-sectional study design was conducted at Muhimbili National Hospital (MNH) in Dar es salaam, Tanzania. MNH is a tertiary hospital with 1500 bed capacity. The diabetic clinic is one of the clinics in the hospital, it operates 3 days a week. About 50–70 patients are consulted by doctors on each clinic appointment day. The clinic receives both insured and non-insured diabetes patients from within and outside of Dar es Salaam. The hospital has routine diabetic clinics every Tuesday, Wednesday, and Thursday. Overall, there are 5 Endocrinologists, 1 General practitioner, 9 nurses and 1 medical record officer who run the clinic. The clinic provides routine diabetic services such as glucose monitoring, vitals, weight and height assessment, medication prescriptions and diet advice sessions. Inclusion criteria for the study were adult patients with type 2 diabetes mellitus who have attended the clinic for at least 3 months who consented to participate in the study. Pregnant women and those categorized as having low cardiovascular risk (less than 40 years with no ASCVD risk) were excluded from the study.

Sample Size Calculation

Sample size for this study was conservatively calculated from the formula for cross-sectional design;²² considering that the outcome was binary, an estimate prevalence of 50% with a margin of error set arbitrarily at 5%, the upper limit of the sample size was estimated at 384 participants.

Sampling Technique and Study Procedures

Convenience sampling technique with consecutive enrolment of study participants was used whereby all consenting adult DM patients attending the Muhimbili diabetic clinic were recruited into the study. The recruitment process was conducted by recruiting eligible participants during the study period. Participants’ unique registration codes were also recorded and cross-checked to ensure that none of the participants was recruited more than once.

Study Variables

Dependent variables for the study included statin prescription patterns. The patterns are categorized as statin benefit group versus statin intensity as follows:¹² (a) For patients with diabetes aged <40 years without additional ASCVD risk factors (low risk category) – none; (b) For patients with diabetes aged ≥40 without ASCVD (moderate risk category) – moderate intensity statins; (c) All ages with diabetes and ASCVD (high risk category) – high intensity statins; (d) Age ≥40 years with multiple (≥2 traditional cardiovascular risk factors) (high risk category) – high intensity statins.

Independent variables included sociodemographic characteristics and clinical and laboratory parameters. Demographic characteristics included: age, gender, and education level, insurance status. Clinical parameters included: (a) traditional cardiovascular risk factors: smoking history, duration of diabetes mellitus, obesity and hypertension; (b) atherosclerotic cardiovascular diseases: acute coronary syndrome, peripheral arterial disease and cerebral vascular events; (c) chronic diabetic complications: neuropathy, retinopathy and nephropathy; (d) cardiovascular risk stratification including moderate and high risk strata (where moderate risk <2, while high risk ≥2 traditional risk factors).²³ Laboratory records included: low-density lipoproteins (LDL) cholesterol, protein in urine, creatinine, glycosylated hemoglobin (HbAIc). Estimation of kidney function (eGFR) – from Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula.

We sought to establish the proportion of participants who belonged to the “very high risk” category, this being defined as documented/established to have any of the following conditions: acute coronary syndrome, coronary artery disease, peripheral arterial disease, chronic kidney disease (nephropathy), and cerebrovascular events.^24–26

Data Collection

The investigator interviewed the patients and recorded laboratory results from the electronic system for all the recruited participants. Statin prescription was also confirmed from the electronic system. An interviewer-based structured case report form (questionnaire) was administered to consented patients or caretakers for those who are unable to communicate. Information collected included sociodemographic characteristics such as age, gender, marital status, education level, and insurance status. Medical information collected included current status of risk factors such as the history of hypertension, smoking, diabetes duration, acute coronary syndromes (ACS), arterial revascularization, stroke, peripheral artery disease. Use and dosage of statins whether low, moderate or high intensity was recorded. Diabetic and other medications in use were also recorded.

Clinical Measurements

Weight was measured on a SECCA weighing scale without shoes and recorded to the nearest 0.5 kg. Height was measured using a height measuring rod and recorded to the nearest 0.5 cm. Body mass index (BMI) was calculated as weight in kilograms (kg) divided by height in meters (m) squared expressed as kg/m². The value obtained was classified as underweight (˂18.5 kg/m²), normal (18.5–24.9 kg/m²), overweight (25–29.5 kg/m²) or obese (˃30 kg/m²).²⁷ Blood pressure was measured by placing a cuff 2 cm above the elbow crease on the bare arm with the midline of the bladder on the brachial artery. Then obliteration pressure was obtained and cuff was inflated 20–30 mmHg beyond the obliteration pressure. Cuff was deflated and the first and fifth Korotkoff sounds were recorded as systolic and diastolic blood pressures in mmHg respectively.²⁸ Hypertension was defined as SBP of at least 140 mmHg and DBP of at least 90 mmHg.²⁹ Alternatively, hypertension was defined as a documented history of hypertension with the current use of antihypertensive medications regardless of blood pressure level at the time of recruitment. Results were recorded on the clinical characteristics case report form.

Laboratory Investigations

Each patient’s serum cholesterol levels, proteinuria, creatinine and HbAIC levels were documented provided that these investigations were obtained in the previous visits. HbA1c control target of 6.5% according to American Diabetes Association was used.³⁰ Low density lipoprotein (LDL) cholesterol was categorized as normal if it fell within the hospital’s chemistry laboratory reference range of 2.6–3.6 mmol/L.

Statistical Analysis

Data were entered and analysed using SPSS version 23.0 and exported to R software version 4.1.2 for analysis. Categorical data were summarized as frequencies and percentages whereas continuous variables were summarized using mean or median as appropriate depending on distribution. Groups were compared for differences using either Chi-square test or Fisher’s exact test. All variables with p <0.2 in the bivariate analysis were entered into the multivariate analysis model. Furthermore, a backward elimination approach was used to select which variables would be included in the final model using the Akaike information criterion (AIC);³¹ p <0.05 was considered statistically significant.

Ethical Considerations

Ethical clearance was obtained from the Muhimbili University of Health and Allied Sciences Research and Publication Committee (Ref: MUHAS-REC-06-2020-284). Permission to conduct the research was sought from MNH administration. Eligible patients were recruited only after verbal or written consent from themselves or their next of kin, for patients who could not provide verbal or written consent. All the principles of the Declaration of Helsinki were adhered to during the data collection process.

Results

In a 3-month study from September to November 2020, 400 participants were approached; 5 participants had low risk for cardiovascular disease (not requiring statin therapy), hence were excluded, the remaining 395 were studied and analyzed, as shown in Figure 1.

Figure 1 CONSORT diagram to show the recruitment process of participants. Note: Adapted from Schulz KF, Altman DG, Moher D, for the CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. Br Med J. 2010;340:c332. doi: 10.1136/bmj.c332. Creative Commons.48

The mean (±SD) age of the study participants was 58.1±10.3 years, out of which 371 (93.9%) belonged to the age group ≥40 years. Two-thirds (241; 61.0%) of the participants were females. Slightly above half of the participants (208; 52.7%) had elementary to primary education. About two-thirds (257; 69.4%) of patients had health insurance coverage. As per BMI (177; 44.8%) were overweight. About half the participants (192; 48.6%) had high blood pressure of ≥140/90 mmHg. The mean (±SD) duration of diabetes was 10.2 (±7.6) years with half of them having ≥ 10 years of the disease.

In the cohort of 395 participants, 307 (77.7%) participants had hypertension, 8 (2%) had CAD, 12 (3%) had PAD and 27 (6.8%) had CVEs. Documented cardiovascular complications, namely neuropathy, retinopathy and nephropathy were found in 309 (78.3%), 234 (59.2%), 121 (30.6%) of study participants, respectively.

For participants who had their LDL recorded, 59 (42.7%) had normal LDL while 82 (20.8%) had high values. Also for patients whose HbAIc was recorded, 44 (27.3%) reached the control target of <6.5%. Proteinuria was found in 104 (26.3%) and decline in kidney function as expressed by estimated glomerular filtration rate was found in 81 (32.5%). As per atherosclerotic risk stratification, the majority of participants (348; 88.1%) had high risk followed by moderate risk (47; 11.9%). The sociodemographic and clinical characteristics of the participants are described in Table 1.

Table 1 Sociodemographic and Clinical Characteristics of type 2 Diabetes Participants (n = 395)

The most frequently prescribed type of statin was atorvastatin (95%) at the dosage of 20 mg, followed by rosuvastatin (5%) at the dosage of 10 mg (Figure 2).

Figure 2 Prescribed statins among study participants.

Statin Prescription Patterns in Relation to CVD Risk Stratification

In this study, 395 (98.8%) participants were eligible for statins prescription however only 187 (47.3%) were prescribed with statins. About 47 (11.9%) had moderate risk, and among them 21 (44.7%) received moderate intensity statin. The majority of the study participants had high risk (348; 88.1%) among which 166 (47.7%) received moderate intensity statins and the remaining 182 (52.3%) did not receive statins. Moderate intensity statin was the only pattern found in this study with none of the participants being on either low intensity or high intensity statin prescriptions (Figure 3).

Figure 3 Prescription of statins in relation to risk stratification among type 2 diabetes participants.

Participants in the “very high risk” cardiovascular category accounted for 133/395 (33.7%). Of these participants, 82/133 (61.7%) were on moderate intensity statins prescription (Table not shown).

Factors Associated with Statin Prescription Patterns

Bivariate analysis revealed that participants aged ≥40 years were significantly more often prescribed with moderate intensity statin than those of <40 years (49.9% versus 8.3%, p< 0.001).

Moderate intensity statin prescription pattern was seen more in participants with health insurance than those without (175; 63.6% versus 10.0%, p< 0.001).

Participants with diabetes duration of ≥10 years were significantly more often prescribed moderate intensity statin than those with less than <10 years diabetes duration (54.7% versus 39.7%, p = 0.003).

Individuals with history of hypertension were significantly more often prescribed with moderate intensity statins than non-hypertensive individuals (52.2% versus 30.0%, p<0.001).

Participants who had coronary arterial disease were significantly more often prescribed with moderate intensity statins therapy compared with those without coronary artery disease (87.5% versus 46.5%, p = 0.023).

Participants with high levels of LDL were not significantly prescribed with moderate intensity statins compared with those with normal LDL levels (68.3% versus 64.4%, p = 0.763). Participants with positive protein in urine were not significantly prescribed moderate intensity statins than those with negative protein in urine (57.0% versus 60.7%, p = 0.634).

Moreover, cardiovascular complications among participants such as neuropathy, retinopathy and nephropathy were seen in 78.3%, 59.2%, and 30.6%, respectively, and moderate intensity statins were prescribed more in nephropathy (62%), followed by retinopathy (50.9%) and neuropathy (48.5%).

Of the cardiovascular complications, cerebral vascular event was the most frequent documented complication (6.8%), followed by peripheral arterial disease (3.0%) and coronary arterial disease (2.0%). Among the participants with documented complications, moderate intensity statins were prescribed in at least half (50%) for each of the individual complications; with the only exception being among participants with neuropathy whom only 48.5% were on moderate intensity statin prescription (Table 2). However, gender, education level, prescriber’s title, history of cigarette smoking, reduced eGFR <60 mL/min/1.73 m² and HbAIc ≥ 6.5% were not associated with moderate intensity statin prescription (Table 2).

Table 2 Factors Associated with Statin Prescription Patterns Among Type 2 Diabetes Participants (N = 395)

The multivariate logistic regression examined associations between statins prescription patterns and various factors. Age was not associated with statins prescription patterns (OR: 0.203; 95% CI: 0.02–1.87). Having medical insurance increased the likelihood of being on statins prescriptions compared with not being insured by 94.4% insurance coverage (OR: 0.056; 95% CI: 0.03–0.12). Furthermore, being hypertensive increased the likelihood of being on statins prescription by 74.1% (OR: 0.259; 95% CI: 0.12–0.54). Diabetes duration, body mass index, education level and proteinuria were not associated with moderate intensity statin prescription after adjustment for the other variables in the model (Table 3).

Table 3 Multivariate Logistic Regression on Factors Associated with Statin Prescription Patterns

Discussion

The current study sought to provide data on statin prescription patterns for the prevention of CVD among participants with type 2 diabetes mellitus. Prescription patterns were evaluated in relation to risk stratification in patients with T2DM. It was found that statins were prescribed in about half of the participants with T2DM who were eligible for prescription. Findings of this study were similar to those found by Mwita et al.,³² Demoz et al.³³ and Gupta et al.²³ and in which statin prescriptions were 45.5% and 55.7%, respectively. The explanation for such a low prescription rate could be due to the following reasons: firstly, lack of local guidelines, secondly, inadequate adherence to already recommended international guidelines,^33,34 and thirdly, limited access to medications and lipid tests due to poor socioeconomic status.

Of those patients who received statins (n = 187), moderate intensity statins prescription was the only pattern found in 47.3%, similarly to 46.1% in the study done by Dermoz et al.³² Likewise, almost half of patients (166; 47.7%) in a high risk group (n = 348) were prescribed with moderate-dose statins, contrary to the recommendations^12,35 while more than half in this group did not receive statins despite being at high risk.

In this study, atorvastatin was found to be the most commonly prescribed type of statin, followed by rosuvastatin. Similarly, in a study done by Gupta et al.,²³ atorvastatin was the most commonly prescribed type of statin (74.1%), followed by rosuvastatin (29.2%). In contrast to this, Demoz et al.³³ found that simvastatin (37.2%) was the most commonly prescribed statin, followed by atorvastatin (32.8%) and rosuvastatin (15.6%). Reasons for discrepancies could be availability, cost, and physician preference.^36,37

With regards to cholesterol levels it was found that having high levels of low density lipoprotein was not associated with statins prescription pattern. Our findings are in contrast to the observation from previous studies that revealed clinicians tend to prescribe statins based on the baseline LDL-C levels rather than to the patients’ overall CVD risk profile as described in clinical guidelines.^38,39 Although there is a lack of local guidelines, MNH diabetes clinic has adopted ADA guidelines which recommend statins along with lifestyle changes regardless of cholesterol levels for all patients with diabetes aged >40 years with or without CVD.¹² These findings may suggest a need for deliberate efforts for improving the understanding and implementation of the adopted guidelines.

Concerning insurance coverage, those who had medical insurance had very high odds of being prescribed with statins, as observed in other studies.^40–42 This can be explained by the easy access of these drugs in insurance packages, hence facilitating prescription. On the other hand, even among those who are medically insured and eligible, some were not on statin prescription. This might as well have been caused by an absence of guidelines in the hospital to guide statin prescription.

Likewise, participants who had history of hypertension as a comorbid condition were significantly more often prescribed with statins compared with those without hypertension. This association has also been reported by various other studies.^43,44 Therefore having a comorbid condition such as hypertension along with diabetes should be regarded as a high risk for developing cardiovascular diseases, hence this should trigger intervention by clinicians in order to provide the best care to these patients before they succumb to cardiovascular complications.

In our study, proteinuria was not associated with statins prescription pattern. This finding is in contrast to that of Berthold et al. who reported that patients with proteinuria had increased odds of statin prescription in Germany.³⁴ Since proteinuria is a marker of renal disease, it is disturbing to see participants with proteinuria in this study received lesser statin prescriptions. We highly encourage clinicians to recognize proteinuria as a predictor of CVD and to increase statins prescription in this group of participants.

Interestingly, the present study found that of those participants who received statin therapy, only 38/94 (40.4%) (Table 2) were found to have normal range LDL-cholesterol level. Certainly, only one third (35.7%) of patients in this study had a detailed lipid profile. This indicates noncompliance with the standard guidelines about monitoring of lipid profile, i.e., that all patients with type 2 diabetes should be tested at least annually.³⁵ This might also be the reason for the low percentage of patients with normal range cholesterol levels. Therefore, there is a need to emphasize periodic monitoring of patients’ lipid levels in optimizing the utilization of statins in response to the CVD risks factors. Such an initiative will ensure that statin prescription decisions are, at least in part, based on the risk assessment and lipid profile.

Multivariate regression analysis revealed that there was no significant association between diabetes duration and statin prescription. In contrast some studies and clinical guidelines have reported association of longer diabetes duration and tendency to statin prescription.^31,45 The difference here may be because almost half of the patients (49%) had less than 10 years diabetes duration from the time of diagnosis.

Smoking was also not found to be associated with statin prescription in this study. This lack of association was also found in other studies.³¹ Our finding is in contrast to other studies which showed that patients with history of smoking had increased odds of being prescribed with statins.⁴⁶ The overall low smoking and alcohol consumption habits in the Tanzania population could contribute to this effect.²⁶ Also, one of the lifestyle modifications in diabetes mellitus is to limit cigarette smoking.

Increase in age and BMI were also not associated with statin prescription in this study; this was contrary to other studies^46,47 where the odds of being prescribed statins were seen in more elderly and in overweight/obese patients. This could be attributed to differences in population representation or the fact that prescribers fear to increase the burden of pills to these high-risk groups to avoid adverse effects as the majority have also been prescribed with other medications.

The findings of this study should be interpreted against the background of several limitations. These include lack of standard risk stratification tool and lack of systematic collected data on microvascular complications (especially renal disease). There is also a possibility of recall bias in participants to some of the disease presentations that could have resulted in inaccurate estimation of their comorbid conditions. Moreover, a qualitative analysis to determine causes of low statin prescriptions by clinicians was not performed. On the other hand, cardiovascular risk categories are derived from international societies and not from Tanzania or regional-specific guidelines. Therefore, one of the limitations is that these CV risk scores and categories might be different in patients in Tanzania (more research is needed to define population-specific risk scores/charts). Categorization of participants into very high risk was based on documented complications, this is likely to be an underestimate as we did not actively assess the participants to obtain any undocumented complications. The use of convenience sampling hinders external generalization of the findings of this study to the general population of diabetes patients in Tanzania. This is the first hospital-based cross-sectional study of its kind in a tertiary hospital in Tanzania that assessed statin prescription patterns and its associated factors. The study has helped to identify the gaps in statin prescribing practice among clinicians and raise awareness for development of local guidelines in order to guide treatment.

Conclusion

This study showed suboptimal statin prescription among type 2 diabetes participants attending a diabetes clinic at Muhimbili National Hospital. Moderate intensity statins prescription was the only pattern found among diabetes participants. Strategies to optimize prescriptions at the clinic should focus on exploring reasons for low statin prescription, clinicians’ awareness of recommended guidelines, continuing medical education as well as time to time prescription audits and dissemination of results to improve quality of preventive care among patients with type 2 diabetes mellitus.