In this real-world study based on a large US EMR database, our study cohort of patients initiating BI from OAD(s) had a mean HbA1c of 9.1%, and approximately 49% had an HbA1c > 9.0% at baseline. This underscores that patients may remain on OAD(s) for extended periods, despite being significantly above recommended glycemic control targets. Further, it indicates that the patients selected for RCTs of BI are generally in better glycemic control than patients who initiate BI in real life; in fact, the average baseline HbA1c was higher in our study than in 36 of 44 RCTs of BI included in a systematic review . However, the results presented here are in line with previous real-world studies in which HbA1c at initiation of BI was well above the recommended target, showing that early intensification to the recommended BI treatment is uncommon in routine clinical care [11, 17, 18]. Factors underlying such delays are complex [20,21,22] and may reflect physicians’ perceptions of patients’ lifestyles and non-adherence to existing medications; and patients’ concerns about out-of-pocket expenses, and reluctance to consider BI, as well as fear of hypoglycemia and concern about weight gain [11, 23].
After initiation of BI (with or without concomitant OADs), there was a robust drop in mean HbA1c of about 1.5% over the first 6 months, but with no additional change from baseline over subsequent quarters. Few real-world studies have examined both HbA1c change after 3–6 months of BI treatment and after long-term follow-up [11,12,13,14,15,16,17,18,, 12, 17]. A US retrospective analysis of patients with T2D found that in 4387 new initiators of BI with an average HbA1c of 9.5% at baseline, the HbA1c decrease of 1.3% at 3 months was the same as the decrease from baseline at 12 months . Similarly, a retrospective longitudinal analysis of EMR from five European countries and the USA showed a similar trend, with the majority of the mean HbA1c decline within the initial 6 months, and no further improvement after 12 months . As these data represent a cross-sectional analysis of a cohort, no consideration is given to the impact of patients who may achieve control initially, and subsequently return to an HbA1c ≥ 7% again.
To better mimic the situation of the healthcare practitioner (HCP), we estimated the conditional probability of reaching glycemic control if continuing BI treatment, beginning in the second quarter post-BI initiation until 2 years, given the condition that glycemic control has not been achieved up to the beginning of the specific quarter. This provides a potentially useful perspective for prescribers who initiate their patients on BI and need to estimate the likelihood over time that a patient will achieve glycemic goals. Our study results suggest that if a patient has been using the same BI regimen for 6 or 12 months and has not yet reached their glycemic target, the likelihood of achieving success on the same regimen is low, which should prompt consideration of treatment modification or intensification. While the estimated probabilities can be affected by the relative robustness of the EMR database and need further confirmation, the overall decreasing trend and the very low probabilities of reaching HbA1c targets after the first year highlight the need for paying attention to patients who have failed to achieve glycemic targets in the first year following BI initiation. There is often a delay in treatment intensification despite persistently elevated glucose levels [13, 14, 24]. In a real-world observational study published in 2016, the median time to treatment intensification in patients with elevated HbA1c following BI initiation was 4.3 years . Multiple factors may contribute to such delays  including concerns related to some treatment options beyond BI. Basal-bolus and premix insulin regimens are potential options, but hypoglycemia, weight gain, and the need to take multiple daily injections can be major concerns [26,27,28]. With the recent development of medications such as GLP-1 RAs, fixed-ratio combinations of BIs and GLP-1 RAs, and SGLT2 inhibitors, which do not increase the risk of hypoglycemia or induce weight gain [29,30,31], prescribers now have more options to consider.
The cumulative probability of reaching glycemic control over time (Fig. 3) revealed that about 38% of patients reached glycemic control in the first 12 months but only about 8% more did so in the second year. These results are also in line with other real-world research [11, 17, 18] which has generally found little further increase in rates of patients with HbA1c below the glycemic target with extended treatment. The previously mentioned US-based retrospective analysis  found an increase in patients below target for new initiators of BI, from 11% at baseline to 27% at 3 months after BI initiation; however, this fell to 25% after 12 months. The aforementioned retrospective longitudinal EMR analysis from five European countries and the USA  found that 20.9% had an HbA1c ≤ 7% at 3 months after BI initiation; by 24 months post-initiation this had only increased to 27.8%. An additional US EMR analysis found that 44% of its cohort achieved HbA1c ≤ 7% within 1 year after BI initiation, with 58% reaching this goal over the entire 2.5-year follow-up . Differences in the patient group (39.2% of this cohort had HbA1c > 9% at baseline, compared with 48.8% in our study) and the specific glycemic goal may play a role in the results.
In the current study, the subcohorts of patients taking no or one concomitant OAD at BI initiation had a slightly higher rate of achieving HbA1c < 7% than those taking two or more OADs, which might reflect differences in stage of disease progression. While the Explorys database cannot definitively provide disease duration for the majority of patients, it is possible that those patients who continued treatment with two or more OADs while initiating BI were considered more progressed and therefore had lower residual beta cell function, which could impact glycemic response to BI treatment. Corresponding broadly to results seen here, a retrospective database analysis of 1830 patients found greater achievement of glycemic targets among those patients taking fewer OADs at baseline (38.2%, 26.7%, and 19.6% for patients taking one, two, and at least three OADs, respectively; p < 0.0001) . There could have been other contributing factors; for example, patients on more medications may have more challenges to be compliant with the prescriptions.
Our study has some important limitations. In the USA, patients change insurance coverage and/or HCPs from time to time because of employer decisions, job changes, relocations, personal choices, etc.; the Health Insurance Portability and Accountability Act rules typically preclude the database from linking multiple records from different sources using personal information. The EMR data collected in the Explorys database provides only a snapshot of patients’ medical histories, and cannot provide the rigor and completeness of data that is typically expected from a prospective longitudinal clinical trial. Laboratory variables are assessed for clinical practice instead of research purposes, and are from multiple sources, obtained as per local clinical practices. HbA1c measurements were not consistently performed every 3 months in many patients, reducing the assessable number of patients included in the analysis. In addition, it is possible that HbA1c measurements may have occurred outside of the data capture infrastructure, and it cannot be guaranteed that those patients with missing HbA1c records behave the same as those with available HbA1c results. The paucity of fasting plasma glucose data and of dosing information in the database are other limitations which make it difficult to evaluate titration practice after BI initiation, or any possible relationship between insulin dose and glycemic response. Given the difficulty of achieving glycemic goals reported in real-world literature, clinicians may find value in understanding which factors predict achievement of an HbA1c target. For example, assessing durability of glycemic control in those treated with BI, along with treatment patterns after BI initiation, can add important further information on factors associated with significantly higher rates of reaching the goal of HbA1c ≤ 7% . We did not do this analysis in our current study because of limitations of the data, but it is our plan to do so in ongoing studies.