Date Published: June 25, 2019
Publisher: Public Library of Science
Author(s): Bastian Rosner, Andres Roman-Urrestarazu, Wisit Cheungpasitporn.
In 2017, more than 1.1 million children were living with type 1 diabetes mellitus (T1DM) globally. The goal in paediatric diabetes therapy is reaching optimal glycaemic control as early as possible in order to avoid complications and early mortality without compromising the quality of life (QoL) of children. Several different insulin regimens are available for T1DM patients to reach this goal.
This review set out to analyse whether continuous subcutaneous insulin infusion (CSII) regimens are superior to multiple daily injection (MDI) therapy in T1DM youth regarding QoL. Additionally, it assessed glycaemic control and adverse events as secondary outcomes and discussed potential future public health implications and justifications for using CSII as a first-line therapy in diabetic youth.
A systematic review and random effects meta-analysis was performed on studies investigating the association between QoL and diabetes treatment regimen. Differences in adverse event rates between groups were analysed using a Mann-Whitney U test. Lastly, differences in glycaemic control were assessed using a random effects meta-analysis.
QoL and glycaemic control was significantly better in CSII subjects at baseline and follow-up. No significant differences in adverse events were found between study groups. No significant changes over time could be shown for either QoL or glycaemic control.
CSII proved to provide similar or slightly better outcomes in all analysed fields. This is consistent with previous research. However, to make credible recommendations, better-designed studies are needed to investigate the impact of CSII in children.
Diabetes mellitus (DM) is one of the top ten causes of global mortality, having killed 1.6 million people in 2016 alone.[1–3] DM describes a cluster of metabolic diseases, rather than a single illness, that are characterised by chronic hyperglycaemia. The American Diabetes Association (ADA) classifies DM into four general categories with the most common ones being type 2 diabetes mellitus (T2DM) and type 1 diabetes mellitus (T1DM) following in second place.[4, 5] It is estimated that more than 96,000 children under the age of fifteen are diagnosed with T1DM annually, whilst there are 1.1 million children and adolescents below 20 years living with T1DM globally. There are considerable regional differences in the prevalence of T1DM with more than one quarter (28.4%) of paediatric patients living in Europe and more than one fifth (21.5%) living in North America and the Caribbean. The highest incidence of T1DM can be seen in the United States (US), India and Brazil. Complications in T1DM are relatively frequent and can be divided into acute (e.g. diabetic ketoacidosis, infection) and chronic (macro- and microangiopathy). In addition to being a global health problem due to its multiple short and long-term complications, diabetes and related conditions account for an enormous economic burden throughout the world. This burden is expected to continue growing with a projected expenditure of 776 billion US-Dollar by 2045 for adult patients only. T1DM’s physiopathology is primarily due to β-cell destruction and absolute insulin deficiency. Thus, the therapeutic goal for T1DM patients is defined as reaching optimal glycaemic control as early as possible to avoid acute and chronic complications without compromising the quality of life (QoL) and wellbeing of children, their parents or caregivers. The only way to reach this goal for patients with T1DM is—additionally to behavioural interventions—the uninterrupted supply of insulin. Insulin regimens available for T1DM patients can be divided into three groups: multiple daily injection basal-bolus insulin regimens (MDI), mixed (biphasic) regimens and continuous subcutaneous insulin infusion regimens (CSII, insulin pump). Despite MDI still being the first-line therapy in many regions around the world[9–11] CSII is gaining popularity among paediatric patients. This can be explained to some extent by slightly better metabolic control and less acute complications through CSII[12–14] but might also be influenced by other factors not yet fully understood. With CSII being much more expensive than MDI–treatment cost would increase by 50% if all T1DM patients used CSII–methodologically well-conducted studies are needed to prove its superiority over MDI and to justify it as a first-line choice.
T1DM is the most common type of diabetes in children. Although T2DM is becoming more common in children and adolescents in some regions around the world, sufficient and reliable data on T2DM in childhood is sparse which makes an analysis of its global health impact difficult. Whereas T1DM can only be treated by insulin injections, there are multiple options for T2DM. Thus, this work will focus on paediatric patients with T1DM only. The definition of childhood provided by the World Health Organization (WHO) was used to set an age threshold for study inclusion criteria. According to the WHO, an adolescent is a person “10 to 19 years inclusive” and a child “is a person 19 years or younger”. Therefore, studies including participants older than 19 years were excluded.
A systematic literature search was performed on 12 December 2018 using PubMed, Web of Science and the Cochrane Library as primary data sources (Table 2 and Fig 1). Studies were selected upon meeting the eligibility criteria stated in Table 1. Additionally, Google.com was searched for grey literature and supplementary data sources. Also, reference lists of included studies and past reviews were screened for more relevant articles. Two levels of screening by two independent researchers (B. Rosner and A. Roman-Urrestarazu) were used on all citations. Our electronic search yielded 1,733 articles (Fig 1). We reviewed the titles and abstracts and eliminated any articles that clearly fell outside our inclusion/exclusion criteria. If there was any doubt, the article was retained for the next level of scrutiny. This process yielded 124 articles. Two authors examined each article’s title and abstract more closely and, if needed, examined the full text of each article and made independent judgments as to whether the article met inclusion and exclusion criteria. Disagreements were resolved by face-to-face discussion, leading to a consensus judgement. Fifteen articles met our inclusion and exclusion criteria.
Fifteen eligible studies could be identified and were included in the analysis of this paper with an agreement percentage between raters of 89.5% and Cohen’s Kappa: 0.604 (Table 3). No additional data was included through screening of references and grey literature. The overall quality of studies was poor (S1 Table). Eligible papers were published between 2003 and 2018, sample sizes ranged from 16 to 700 and all studies included male and female patients in their analyses. Study centres were based in the US[31–34], Germany[8, 35], Denmark[36, 37], Italy, the Netherlands, Hungary, England and Wales, and Israel[41–43] (with one of the Israeli studies additionally including patients from a study site in Slovenia). Research methodology varied substantially, particularly because of different study designs used. Six studies were randomised controlled trials (RCT)[8, 15, 33, 39, 42, 43] (of which two applied a crossover design[42, 43]), another six were cross-sectional studies (CSS)[31, 32, 34, 36, 38, 40] and one each were a clinical trial (CT), a crossover CT and a prospective observational study. The age of study participants ranged from 0.6 to 19 years (Table 3). Three of the studies[8, 15, 35] stated results for different age groups with Mueller-Godeffroy et al.[8, 35] and Blair et al. reporting for patient cohorts younger than 8 years, 8–11 years, 12–16 years and cohorts younger than 5 years, 5–11 years, 12–15 years respectively. Inclusion criteria typically consisted of all patients being diagnosed with T1DM, being currently treated with MDI (RCT) or being either treated with CSII or MDI (CSS) before the study. Only one paper included newly diagnosed T1DM patients with no prior treatment. Given consent by patients and their parents as well as the absence of major comorbidities were additional eligibility criteria. All studies reported HbA1c and QoL as outcome measures, nine studies[15, 33, 35, 37–39, 41–43] stated the number of occurred adverse events within each study group (Table 4). Other reported secondary outcomes—which are not the focus of this review—were cardiorespiratory fitness, parenting stress, treatment satisfaction, insulin dose, BMI, hypoglycaemia fear and cost-effectiveness.
The current meta-analysis could show significant differences in QoL between paediatric CSII and MDI users at follow-up. However, QoL was better in CSII subjects at baseline, too, which is a bias potentially mitigating the validity of the performed analysis. No significant change in QoL over time could be observed. The evidence suggested no significant differences in adverse event incidence between treatment groups. Severe hypoglycaemia incidence was higher in MDI subjects whereas ketoacidosis incidence was shown to be higher in CSII subjects. A significant difference in glycaemic control could be shown between treatment groups both, at baseline and follow-up, with CSII yielding lower HbA1c values at both points in time respectively. Despite the difference being more prominent at follow-up, no significant change could be shown over time.
This paper set out to analyse whether CSII regimens are superior to MDI therapy in T1DM youth regarding HRQOL. Additionally, it assessed glycaemic control and adverse events as secondary outcomes and discussed potential future public health implications and justifications for using CSII as a first-line therapy in children and adolescents. Despite the stated limitations and the fact that no considerable difference in QoL between treatment groups could be shown over time, CSII proved to provide similar or slightly better outcomes in all analysed fields. This is consistent with previous research. However, to make credible and reliable recommendations, bigger, better-powered and better-designed studies are needed to investigate the impact of CSII in children. Poor methodology, small samples and short follow-up times constrain the ability to assess the association between QoL and CSII to the full extent.