Less than half of the statin-eligible patients with type 2 diabetes at a specialised diabetes clinic in Botswana received a statin prescription. A longer duration of diabetes, a higher baseline LDL-C and the presence of chronic kidney disease were independently associated with the tendency to prescribe statins.
The under-prescription of statins in this population is a concern since the use of statins appreciably reduces cardiovascular morbidity and mortality in patients with diabetes irrespective of their LDL-C levels [7, 9-13]. Whilst the proportion of patients with diabetes who are prescribed statins varies substantially worldwide; there is a low prescribing of statins both in developing and developed countries [18, 19, 21-23, 49-52]. Encouragingly, the percentage of patients with diabetes who received statins (45.5%) in our study appears appreciably higher than the 3-13% seen in some African countries and consistent with findings from developed countries where between 25% to 73% of patients with diabetes are prescribed statins [18, 19, 22, 49-51]. The proportion of statin prescription in this population is higher than has been reported in some developed countries, including Germany (25%) and the United Kingdom (33%)[18, 49]. While the finding of a comparatively higher statin prescription in this setting than some African countries and some developed countries is encouraging, there is no reason for complacency as more than half of our patients were without CVD protection by statins. Similar to developed countries, one potential explanation for low statin prescribing rates among our patients with diabetes is inadequate adherence to guidelines [22, 52-54]. Whereas there may be a fear of the association of statin therapy with a slightly increased risk of developing diabetes, the benefits of statins in reducing cardiovascular morbidity and mortality among patients with established diabetes should dispel these concerns [7-9, 55]. As mentioned, several epidemiological studies have observed a lower proportion of statin prescribing in patients with diabetes in Africa than in this population [21-23]. In addition to inadequate adherence to guidelines, the main reasons for low statin prescribing in Africa include limited access to these medicines due to their high cost with high co-payment levels, inaccessibility to lipid testing facilities and unavailability of guidelines [22]. The presence of free consultations, tests and medications in Botswana might explain our higher statin prescribing rates than those seen in other African settings without universal health access. Irrespective of the reasons, it is imperative that statins are routinely prescribed to reduce the risk of CVD events in patients with type 2 diabetes [7-13, 15, 56].
Our results of increasing statin prescribing with increasing diabetes duration also agree with previous research findings [57]. This finding is reassuring as a longer duration of diabetes leads to an increased risk of CVD. For this reason, guidelines recommend statins for patients with diabetes for more than ten years [5, 6]. Although the median diabetes duration was seven years in our participants, the association between statin prescribing and diabetes duration was still apparent. The finding may suggest that clinicians correctly recognise a longer duration of diabetes as an indication for statin therapy. As it may take time for the transmission of information between clinicians and patients, the acceptance of new medications is likely easier as the diabetes duration increases [58].
Another finding in our study was that the presence of CKD increased the likelihood of statin prescribing. This finding is also encouraging as statins reduce mortality by up to 36% in patients with kidney failure [5, 6, 59, 60]. The finding is also consistent with SEMDSA guideline recommendations of a statin for every patient with diabetes and CKD [6]. While albuminuria is a marker of renal disease, participants with proteinuria did not receive statin prescriptions. We can postulate that clinicians do not recognise proteinuria as a predictor of CVD and an indication for statins in patients with diabetes. This finding warrants further investigation as it is in contrast with Berthold et al., who reported increased odds of statin prescribing in type 2 diabetes patients with proteinuria in Germany [18].
Our findings that a high baseline LDL-C increased the likelihood of statin prescribing agreed with those of Berthold at al. that showed an 11% increase in statin prescribing rates for every 0.26 mmol/L increase in LDL-C [18]. This finding also confirms the observation from previous studies that prescribers tend to respond more to the pre-treatment LDL-C value than to the patients’ overall CVD risk profile as described in clinical guidelines [9, 53]. Although there is a lack of local guidelines, our diabetes clinic adopted the SEMDSA guidelines which recommend statins along with lifestyle changes regardless of cholesterol levels for all patients with diabetes aged > 40 with or without CVD[6]. Our findings that statin prescription was based on LDL-C level may suggest a need for deliberate efforts for improving the understanding and implementation of the adopted guidelines, and we will be taking this further.
In most clinical guidelines, the presence of CVD, CKD, patients age, diabetes and presence of CVD risk factors such as hypertension, albuminuria and cigarette smoking are indicators of prescribing statins among patients with type 2 diabetes[5, 6]. The presence of any of the above factors is associated with an increased risk of CVD. Except for CKD and duration of diabetes, none of the other indications was a predictor of statin prescriptions in this population. Given the high prevalence of hypertension and other indications in this population, most participants would have qualified for statins if guideline recommendations were followed. As our clinic has adopted the SEMDSA guidelines, this finding is a concern and a call for efforts to improve its implementation for the benefit of this high-risk population.
We are aware of a number of limitations of our study. We estimated the baseline LDL-C levels by a 25% adjustment of measured LDL. There was a risk of either overestimation or underestimation of the baseline LDL-C due to possible errors in our assumptions of the dosage and adherence of atorvastatin. Although measured LDL-C results were available for most of the included participants, HDL cholesterol results were mostly missing. Guidelines consider HDL as one of the factors for statin prescribing in patients with diabetes. However, data for all other indications for statin prescribing were available in dataset. We did not document the dosage of statin used in our patients; hence, we are unable to determine whether moderate to high-intensity statins were prescribed as recommended by the guidelines. In addition, we were unable to identify patients with contraindications to warfarin as the information was not available. The study was also performed in one clinic, hence limiting the generalizability of the study findings to other facilities in the country. However, being one of the few specialised diabetes clinics in Botswana, our results likely characterise the ‘finest’ diabetes care in the country. Consequently, the highlighted concerns are likely to be greater in non-specialist healthcare facilities treating patients with type 2 diabetes in Botswana.