Introduction

Type 2 diabetes mellitus (T2D) is a chronic and progressive vasculo-metabolic disease. The seriousness of T2D, coupled with its increasing incidence and prevalence, continues to underscore the need for effective and widely implemented treatment programs [1].

Accordingly, physicians and patients devote considerable energy to managing T2D through lifestyle changes and medical interventions. It has become increasingly clear in recent years that one of the elements to which we must pay particular attention is quality of life (QoL). QoL is a measure of perceived physical and mental well-being, and is reduced in people with T2D compared with the general population [2]. The principal causes of impaired QoL in poorly controlled T2D are health-related complications (such as retinopathy, coronary heart disease, and stroke) [3–5]. However, the lifelong nature of T2D and the commitment needed to manage it assuredly have the potential to create additional pressures. It is perhaps unsurprising that more recent treatment guidelines advocate for QoL as a treatment goal in its own right [1, 6], and it has been suggested that QoL be considered the fifth element of a glycemic pentad (alongside the four key measures of good glycemic control) to be used to guide treatment decisions [7].

Research into the relationship between treatment and QoL in T2D indicates that intensive treatment per se does not impair QoL; in fact, QoL improves along with glycemic control [8]. However, poor QoL can lead to diminished levels of self-care, poor treatment adherence, and thus reduced glycemic control. In patients for whom adherence to medication is less than optimal, it is therefore important to identify factors that can improve QoL [9, 10]. For patients requiring insulin, the range of formulations available is broad; physicians are thus potentially well placed to choose regimens according to their impact on QoL. Conventional insulins, such as regular human insulin and intermediate-acting neutral protamine Hagedorn insulin (NPH), do not replicate the pattern of basal and postprandial endogenous insulin secretion. Insulin analogs were designed to address this limitation [11]. Doubts expressed about their superiority in efficacy and safety terms after their introduction [12] may have been mitigated more recently [13, 14], and both conventional and analog insulins have found a place in clinical practice. Both are additionally used in premix (or biphasic) formulations, in which a proportion of rapid-/short-acting insulin is protaminated to extend its duration of action. Insulin co-formulations and premixed analogs offer comprehensive glycemic control, lowering fasting glucose, postprandial glucose, and HbA1c in a safe and well-tolerated manner. Such characteristics encourage greater treatment adherence and facilitate achievement of optimal therapeutic outcomes [15]. Of premix insulins, the most widely used are insulin analogs, particularly biphasic insulin aspart 30 (BIAsp 30) (NovoMix® 30; Novo Nordisk A/S; NovoLog® Mix 70/30; Novo Nordisk Inc.) and biphasic insulin lispro 25 (LM25) (Humalog® Mix 75/25™; Eli Lilly and Company). Although BIAsp 50, BIAsp 70, and LM50 are available for patients who require more prandial insulin, these formulations are used less frequently. The key benefits of BIAsp 30 and LM25 are favorable efficacy and safety profiles combined with dosing flexibility for people with regular eating patterns, fewer injections than basal–bolus regimens, and ease of intensification [16–23].

This review was undertaken to determine if there is any evidence to suggest that QoL or patient satisfaction can be improved in T2D by switching from a biphasic human insulin (BHI) to a biphasic insulin analog, and if there is any associated evidence for a positive impact on adherence.

Compliance with Ethics Guidelines

This article is based on previously conducted studies and does not involve any new studies of human or animal subjects performed by any of the authors.

Literature Appraisal

PubMed and Google Scholar searches were conducted on 4 April 2016 using variants of the terms “quality of life”, “health related quality of life”, “medication adherence”, “medication noncompliance”, “medication non-adherence”, “medication persistence”, “medication compliance”, “premix insulin”, “premix analog(ue)”, “Humalog Mix 25”, “Humalog Mix 50”, “lispro mix”, “biphasic insulin aspart 30”, “BIAsp 30”, “NovoMix 30”, “NovoLog Mix 70:30”, and “biphasic human insulin”. Insulin degludec/insulin aspart (IDegAsp) is a soluble co-formulation that combines two insulin analogs and was not included in the scope of this review. The initial search returned 167 citations. The current review was based on articles comparing BHI and premix insulin analogs: 11 articles compared QoL outcomes, 13 compared surrogate markers for QoL, and the remaining seven considered adherence and cost-effectiveness. Further articles were added at the discretion of the authors.

Measuring QOL in T2D

QoL is a broad multidimensional concept that includes subjective evaluations of both positive and negative aspects of life with regard to perceived physical and mental well-being, whilst adherence encompasses changes in behavior—taking medication, following a diet, and/or executing lifestyle changes—corresponding with agreed recommendations from a healthcare provider [2, 24, 25]. Factors determining QoL in diabetes are shown in Table 1; measurement of surrogate markers in particular plays an active role in assessing response to therapy [26]. Literature reporting the effect of insulin analogs on QoL specifically is scarce, despite assessment of QoL in diabetes being an increasingly important aspect of care, and research into QoL overall has escalated in the last 20–30 years (a PubMed search using the terms “quality of life” and “diabetes” showed an increase in the number of articles from one to more than 1100 over a 35-year timeline).

Table 1 Determinants of QoL in people with diabetes
Full size table

Researchers have developed useful techniques that have helped to conceptualize and measure these multiple domains and how they relate to each other. These scoring systems have been used for assessing QoL in diabetes, and can be broadly classified as generic (i.e., applicable to all health conditions) or diabetes-specific. No single technique can suit every purpose or application but, when appropriate measures can be selected, robust assessments can be made and have been the subject of review [27]. The validated assessment tools more commonly used are the diabetes treatment satisfaction questionnaire (DTSQ) [28], the EuroQol five dimension (EQ-5D) questionnaire and visual analog scale (EQ-VAS) [29], the diabetes medication satisfaction (DiabMedSat) measure [30], the insulin treatment satisfaction questionnaire (ITSQ) [31], the diabetes health profile (DHP) [32], and the audit of diabetes-dependent QoL (ADDQoL) [33, 34].

Comparing QoL Outcomes for BHI and Premix Insulin Analogs

Premix insulin analogs were designed to improve patient convenience, reducing the number of daily injections and providing flexibility in injection timing by combining rapid- and intermediate-acting components in one formulation. Despite this, QoL measures have not been routinely included in clinical trial designs.

BHI Versus BIAsp 30

Seven articles were retrieved that compared QoL outcomes after patients switched from BHI to BIAsp (Table 2a); all articles reported subgroup analyses from observational studies. Ethnic/geographic variation in baseline QoL scores (while patients were receiving BHI) were apparent in the A1chieve study, although all showed improvements in QoL after switching to BIAsp [35–40]. Patient satisfaction with treatment, meanwhile, improved in the IMPROVE study [41]. Furthermore, Brod and colleagues analyzed the influence of prior and current treatment factors on QoL using data from the 17,488 patients recruited from Asia, Europe, and North America [42]. Analyses were conducted with the whole IMPROVE cohort (i.e., not just those receiving BHI therapy before the switch). As with the subgroup previously receiving BHI, this overall population reported improved treatment satisfaction with the switch to BIAsp 30—an improvement considered clinically meaningful to patients.

Table 2 Articles comparing QoL outcomes for BHI and premix insulin analogs
Full size table

A further two articles were retrieved that compared QoL outcomes with BHI and BIAsp using parallel-group study designs (Table 2b). In the first article, BHI twice daily (BID) was compared with BIAsp 50 or 70 three-times daily (TID) (with BIAsp 30 substituted as the predinner dose according to morning blood glucose [BG] levels in 76% of patients) [43]. No differences were detected between regimens in any of the dimensions evaluated in the two QoL questionnaires. While this may indicate that patients did not consider a TID regimen problematic compared with a BID regimen, it is acknowledged that there may have been insufficient statistical power to detect minor differences in treatment preferences. In the second article, differences in the frequency of low glucose values using either BIAsp 30 or BHI 30 were investigated using a crossover study design [44]. No between-treatment differences were detected in treatment satisfaction or QoL.

BHI Versus LM25/LM50

The literature search retrieved two publications reporting a switch in treatment from BHI (Table 2a). In the first study, patients previously treated with BHI 30 BID were randomized to LM25 (if postprandial plasma glucose [PPG] levels <200 mg/dL) or LM50 (if PPG levels >200 mg/dL) [45]. After 12 weeks, 38 patients reported a preference for injections just before meals (LM), whereas only two preferred injections 30 min before meals (BHI). In the second study, the switch to LM50 did not affect QoL scores; however, the value for convenience did improve with LM50 [46].

The literature search failed to retrieve any relevant publications comparing QoL outcomes with BHI and LM using a parallel-group design.

Comparing Surrogate QoL Outcomes for BHI and Premix Insulin Analogs

Hypoglycemia

Fear of hypoglycemia affects patients’ health-related QoL (HRQoL) and adherence to insulin treatment [47–49]. Low incidence of hypoglycemic events has been shown to reduce anxiety about insulin dosing, thus improving QoL and positively impacting on the ability to maintain glycemic control [50–52]. For premix insulin analogs then, which may provide little benefit over BHI in terms of glycemic control, any relative reduction in hypoglycemia risk may have an important influence on QoL.

BIAsp

Generally, HbA1c levels were not significantly different for BHI and BIAsp regimens in comparative studies (Table 3) [44, 53–55]. The exception was the study conducted by Clements and colleagues, in which HbA1c levels were significantly better with BIAsp than BHI; in this case, however, BHI 30 BID was compared with predominantly BIAsp 50 and 70 TID regimens (BIAsp 30 administered before dinner according to morning BG levels in three-quarters of patients) [43]. The relative risk of overall and minor hypoglycemia in these studies did not consistently favor either regimen, but the risk of nocturnal hypoglycemia was consistently lower with BIAsp, particularly after allowing for differences in dosing frequency between BHI and BIAsp [43, 44, 53, 55]. The risk of major hypoglycemia was generally low with BHI and BIAsp but, on balance, may favor BIAsp.

Table 3 Studies comparing BHI and premix insulin analogs with regard to HbA1c and hypoglycemia
Full size table

Results of individual clinical studies have been largely corroborated by a meta-analysis of nine randomized trials involving 1674 patients, with mean (SD) age of 61.0 (10.6) years and mean (SD) duration of diabetes of 10.9 (7.9) years [20]. BIAsp 30 had a 50% lower rate of nocturnal hypoglycemia than BHI 30 (relative risk: 0.50 [95% confidence interval {CI}: 0.38;0.67]; P < 0.01) (predefined primary endpoint), whereas the rate of daytime hypoglycemia was 24% lower for BHI 30 (1.24 [1.08;1.43]; P < 0.01), and rates of overall hypoglycemia were not significantly different (1.08 [0.94;1.24]; P = NS). The likelihood of major hypoglycemia was significantly lower with BIAsp 30 than BHI 30 (odds ratio: 0.45 [95% CI: 0.22;0.93]; P < 0.05). The authors concluded that BIAsp 30 was associated with a significantly lower rate of nocturnal and major hypoglycemia, but a significantly increased risk of daytime hypoglycemia compared with BHI 30 at a similar HbA1c level.

LM25

Studies comparing efficacy of BHI and LM showed no significant differences in HbA1c levels or rates of overall hypoglycemia (Table 3) [56–58]. One study, however, showed significantly less nocturnal hypoglycemia with LM than BHI, in addition to offering the convenience of dosing closer to mealtime [57].

Hyperglycemic Symptoms and QoL

In a study examining the cross-sectional relationship between glycemic control, hyperglycemic symptoms, and HRQoL in T2D, a relationship between hyperglycemic symptoms and QoL was indicated [59]. A total of 1149 people being treated in primary care in the Netherlands were included in the study. Patients had a mean (SD) age of 68.7 (11.5) years and a median duration of diabetes of 5.3 years. They were asked to complete a Rand-36 questionnaire (appraising physical and emotional health dimensions), as well as indicate which of 10 glycemic symptoms they had experienced in the previous month. Specifically, the hyperglycemic symptoms comprised fatigue, weight loss without clear cause, itching, thirst, increased fluid intake, and increased urine production, and the hypoglycemic symptoms comprised excessive sweating, shaky sensation, dizziness, and sudden appetite, which disappeared when food was taken. Accumulated data following univariate and multivariate analysis indicated that hypoglycemic and hyperglycemic symptoms were strong and independent influences on HRQoL. The influence of hyperglycemic symptoms was stronger; it was also associated with HbA1c, whereas hypoglycemic symptoms were not. The authors concluded that further study was required to validate this outcome, but if the association between hyperglycemic symptoms and HRQoL in the absence of an association between HbA1c and HRQoL is confirmed, hyperglycemic symptoms should be considered a separate objective in diabetes management [59].

Impact of Flexibility on QoL

The flexibility of an insulin regimen can be considered its “ability to be injected at variable times, with variable injection–mealtime gaps, in a dose frequency and quantum determined by shared decision making, with a minimal requirement of glucose monitoring and health professional consultation, with no compromise on safety, efficiency and tolerability” [60]. The validity of flexible dosing time and its impact on QoL have been addressed in numerous publications [18, 61, 62]. A comparison of postprandial versus preprandial dosing of premix analog in 93 older patients (≥65 years of age) with T2D in the USA showed that BIAsp 30 offers a degree of flexibility in terms of injection timing, allowing individual adjustment for meal size and content without significantly affecting efficacy, and that mealtime flexibility adds to treatment convenience [61]. Similar conclusions were reported by Cornell and colleagues when reviewing the need for early insulin treatment in T2D, and assessing QoL benefits associated with insulin analogs compared with human insulins [62]. In a comprehensive comparison of the pharmacology, efficacy, and safety of premixed insulin analogs and premixed human insulin, Garber and colleagues reported that the overall evidence suggests that flexible mealtime dosing with premix insulin analogs can improve adherence and QoL compared with premix human insulin, and hence offers useful advantages for T2D treatment [18].

Adherence and QoL

To what extent does better QoL translate into better adherence? Peyrot and colleagues examined patients’ and physicians’ beliefs regarding insulin therapy and the degree to which patients adhere to insulin regimens; 76% of physicians would have treated closer to target if not for concerns about hypoglycemia [63]. The authors noted that patients using insulin were not entirely satisfied with treatment, mainly because the restrictive regimen made it difficult to take all doses as prescribed. They advised that physicians consider prescribing more flexible regimens and reducing the burden of treatment, preferably minimizing the number of injections and thus the risk of hypoglycemia and consequences of a delayed/missed insulin dose. The results from this survey are consistent with those reported by Rubin and colleagues, in which most patients wanted to reduce the number of daily injections because of the substantial negative impact on QoL [64]. A 26-week, prospective, observational study conducted in Europe investigated safety and efficacy of switching to BIAsp 30, 50, or 70 in 592 patients with T2D previously treated with BHI 30 or 50 [65]. Although there were some shortcomings (the absence of a validated QoL questionnaire and objective measurement of adherence), the authors were prompted to stress the importance of patient education in addressing specific issues and concerns about insulin therapy, and thereby improving adherence to treatment.

Cost-effectiveness

Total global health expenditure for treating people with diabetes in 2015 was USD 673 billion [1]. T2D has, however, been growing at a particularly high rate in low- and middle-income countries, where a disproportionally small amount of the global expenditure on treatment can be expected. Globally, estimated expenditure on diabetes is predicted to rise considerably between 2015 and 2040, as exemplified by the estimates for India (USD 6569 million to USD 11,721 million), Indonesia (USD 1715 million to USD 2829 million), and Saudi Arabia (USD 3994 million to USD 6514 million).

Although the cost-effectiveness of the switch from conventional insulins to analogs has been disputed [66], published data support cost-effectiveness. Ali and colleagues used a validated computer simulation model of diabetes to project long-term outcomes (such as quality-adjusted life expectancy and direct medical costs [based on 2007 Saudi Arabian Riyals, SARs]) based on patient characteristics and treatment effects observed in the Saudi Arabian subgroup of the PRESENT study [67]. Specifically, the cohort comprised 598 patients with poorly controlled T2D, with mean (SD) age of 46.5 (13.7) years and mean (SD) duration of diabetes of 11.3 (6.8) years. Switching to BIAsp 30 from BHI was projected to improve life expectancy by 0.62 years and quality-adjusted life expectancy by 0.92 quality-adjusted life-years (QALYs) while reducing lifetime direct medical costs (53,879 SARs per patient). A similar analysis was conducted using treatment-effect data from 311 patients in the Canadian subgroup of the IMPROVE study (who switched from BHI 30 to BIAsp 30), which was applied to the US setting with 2008 USD [68]. The patient cohort had a mean (SD) age of 64.2 (11.0) years and a mean (SD) duration of diabetes of 16.4 (8.9) years. The authors concluded that switching to an insulin analog may represent a cost-effective treatment option in the USA for patients with T2D experiencing poor glycemic control or hypoglycemia while receiving human premix insulin. BIAsp 30 was projected to improve life expectancy by 0.202 years and quality-adjusted life expectancy by 0.301 QALYs, with lifetime direct medical costs increased by USD 8998. The projected incremental cost-effectiveness ratio was within the range considered to represent good value for money in the US setting at USD 29,870 per QALY gained with BIAsp 30 compared with BHI 30. In a subgroup analysis of 2025 patients from India, Indonesia, and Saudi Arabia from the A1chieve study, switching treatment from BHI, insulin glargine, or NPH to BIAsp 30 was consistently associated with an increase in life expectancy (>0.7 years) and a reduction in all diabetes-related complications [69]. The switches were considered cost-effective (i.e., cost of an additional health benefit gained was not greater than three times the 2011 growth domestic product per capita and costs based on 2013 exchange rates) in all three countries (for switches from BHI, IGlar, and NPH: India 0.26, −0.68, and 0.15, and Saudi Arabia 0.01, −0.21, and −0.07; switch from BHI: Indonesia 1.25). Cost-effectiveness, moreover, was maintained in the 1- and 30-year analyses, indicating short- and long-term benefits.

Discussion

There is increasing evidence that QoL is improved among patients switching from BHI to premix insulin analogs. While this might be attributed in part, directly or indirectly, to factors such as dosing flexibility, which potentially improves adherence and thus glycemic control, it seems plausible that differences in hypoglycemic events (principally, benefits in terms of nocturnal and major events with premix insulin analogs) could be particularly important. Given that patients with a history of recurrent major hypoglycemia were excluded from many trials [12], however, it would be premature to assume any risk reduction and any potential QoL benefit might be observed across all patients. The evidence base currently lacks information about how the newer premix insulin analogs compare with human insulin mixes.

Despite a growing interest in QoL, it must be acknowledged that the evidence directly comparing QoL benefits for BHI and premix insulin analogs is surprisingly limited, with many of the publications derived from a small number of observational studies. Widening the focus to surrogate measures provides additional insights but exacerbates efforts to compare across trials given the heterogeneity in terms of study methods and patient populations. Our conclusions are thus drawn with caution, and we advocate renewed efforts to build QoL endpoints into clinical research, particularly treat-to-target trials, so that we can better understand the interrelated factors affecting QoL and the extent to which it is affected by different treatment decisions. The ongoing MOSAIC study is one such study awaited with interest [69, 70].