Cancer survival has increased over the past 20 years (Richards et al, 2018). This has been achieved through advances in treatment and improved management of treatment-induced side effects (Jeong et al, 2016). Radiotherapy, chemotherapy and general cancer symptoms, commonly require steroids to manage these side effects (Schultz et al, 2018). Steroids increase blood glucose levels, causing new hyperglycaemia in patients with no known history of diabetes, termed steroid-induced diabetes mellitus (SIDM) by the Joint British Diabetes Society (JBDS, 2014; NICE, 2017). Monitoring blood glucose is recommended by the National Institute for Health and Care Excellence (NICE, 2017). However, SIDM is likely to be underdiagnosed, according to Bonaventura et al (2018), due to a lack of both evidence-based monitoring guidance and staff and patient awareness and education, especially in an outpatient setting (Gallo et al, 2018; Storey et al, 2017; JBDS, 2014).
Blood glucose levels do not have to be raised for long before a patient will experience effects and uncontrolled hyperglycaemia can lead to acute problems (Diabetes UK, 2018). For example, such events have affected morbidity and mortality within The Christie NHS Foundation Trust. These potentially avoidable events led to protocol development to monitor blood glucose when commencing steroids within the trust. A local audit found that protocol compliance is poor. This paper explores SIDM in cancer patients, identifies areas that could influence positive changes in current and future practice in its management and discusses their impact on practice in an outpatient setting.
Steroid-induced diabetes mellitus
Steroids are used in various settings in cancer management and treatment side effects, such as nausea, poor appetite and pain flare(Jeong et al, 2016; Ariaans et al, 2015). SIDM is a common and upsetting condition in cancer treatment, and has been recorded at 10% in an inpatient setting (JBS, 2014; Lee et al, 2013). Blood glucose monitoring has been suggested when commencing steroids (Harris et al, 2013; Schultz et al, 2018). Monitoring can lead to prompt intervention thus reducing adverse effects, but there is no clear evidence-based guidance available (Storey et al, 2017; JBDS, 2014). NICE (2017) suggest monitoring blood glucose levels by testing glycated haemoglobin (HbA1c). This reading is taken by measuring glucose bonded to haemoglobin in red blood cells, providing a reflection of blood glucose levels over the previous 8-12 weeks before testing (Diabetes UK, 2018). In practice, capillary blood glucose (CBG) would be a more appropriate test for SIDM, as it assesses current blood glucose levels upon commencement of steroid treatment, which may require immediate intervention (Diabetes UK, 2018). Disparity has been reported in diagnostic blood tests and blood ranges (Storey et al, 2017). This could be improved with the creation of clear clinical guidelines.
The prevalence of SIDM is difficult to establish as there are varying reports, but has been described as 50% in an inpatient setting, and is likely to be underdiagnosed (Storey et al, 2017; Hwangbo and Lee, 2017; Bonaventura and Montecucoo, 2018).
Schultz et al (2018) conducted a prospective observational study of 140 patients with a confirmed cancer diagnosis of metastatic spinal cord compression. Patients with no prior diabetes diagnosis who were receiving radiotherapy of 30Gy in 10 fractions and treated with more than 100mg prednisolone daily were included in this study. The study aim was to establish incidence of SIDM, time to onset and identify any risk factors. A total of 324 patients were recruited from May 2013 to December 2014. Patient selection was not explored by the study authors, and therefore an element of selection bias cannot be eliminated from the study, but was minimised by the collection of baseline information. The patient cohort included patient tumour groups that varied in sex, age and performance status, providing evidence that there was no bias towards selection in these categories. HbA1c, C-peptide, lipids and plasma glucose were taken at the time of inclusion in the study and on the last day of radiotherapy. Identification of SIDM was determined by two random plasma glucose (RPG) ≥ 11.1mmol/L, along with CBG measurements taken 1-4 times a day. The study anticipated a prevalence of SIDM at 15% (±5%), but the actual prevalence was higher than expected, with 56 patients (43%) meeting the criteria for SIDM. There is no description of how anticipated prevalence was calculated. The authors of the study concluded that identifying blood glucose requires regular monitoring and the incidence of SIDM could be significant.
Rowbottom et al (2015) conducted a retrospective analysis of 30 patients with genitourinary cancer treated with steroids, and found 43% of patients had no pre-existing diabetes, and 46% required intervention due to SIDM. There were no guidelines in place when monitoring glucose levels for patients receiving steroids. This study reported three patients had random plasma glucose (RPG) of >27.8mmol/L, two were admitted, and one presented with polyuria and polydipsia. This patient was admitted on two occasions, rapidly declined and died. Three patients had RPG of 15.0-27.8mmol/L, one died due to multiple organ failure and the other two were admitted with hyperglycaemia, requiring the reduction of their steroid dose to achieve manageable blood glucose. 16.7% of patients were admitted when this could have been avoided if there had been guidelines for blood glucose monitoring at home and advice in self-managing medication. The limitations of the study were inconsistent blood glucose monitoring and a small sample size, making it likely that the prevalence of steroid-induced hyperglycaemia was underestimated. The study concludes that SIDM has serious effects that could be minimised with monitoring and intervention.
Daily screening via CBG is transferable to practice as it is readily available, cost-effective, and can be conducted by either the patient or a staff member (Harris et al, 2013; Diabetes UK, 2018). In the Christie NHS Foundation Trust, SIDM protocols stipulate outpatients are referred to a diabetic nurse specialist when they have a CBG of >8.1 mmol/L. The JBDS (2014) recommend intervention with antidiabetic medication for inpatients with a blood glucose level greater than 12mmol/L, but does not provide guidelines for outpatients. This could prompt a practioner go question the CBG values in local SIDM guidelines. Could local guideline levels be increased to ensure there is no unnecessary CBG monitoring, or are there other suitable methods of testing?
Monitoring blood glucose in outpatient practice
Establishing suitable timing and a result range for testing in outpatient practice is challenging, as there are few evidence-based guidelines available (Gallo et al, 2018; Brady et al, 2014). Diabetes is defined as a fasting plasma glucose ≥ 7.0mmol/l (126mg/dl) or 2–h plasma glucose ≥ 11.1mmol/l (200mg/dl) (WHO, 2006). Two tests are required to confirm a diagnosis in asymptomic patients, taken at least a week apart, and one test is required for symptomatic patients. CBG monitoring provides information for intervention (JBDS, 2014). Brady et al (2014) states the Endocrine Society suggests blood glucose monitoring prior to commencement of steroids. Performing a HbA1c test prior to commencing steroid treatment would ensure identification and management of undiagnosed patients, could help identify patients at risk of developing SIDM, and should include plasma glucose testing (Schultz et al, 2018). The HbA1c blood test results can take up to five days to be returned. The HbA1c and plasma glucose tests could be performed easily in an outpatient setting as most oncology patients have a blood test during their first appointment. Sampayo and Tofthagen (2017) state that this useful tool must be exercised with caution, as results are affected by anaemia, which is prevalent in cancer patients, and blood transfusions. In practice, with adequate training and guidance, this can be taken into consideration when reviewing blood results.
In a prospective observation study of 90 patients receiving chemotherapy or radiotherapy treatment, who used steroids for management, Harris et al (2013) defined outpatient hyperglycaemia of RPG=8mmol/L as abnormal, and found 13.3% of patients were in the range that merited a new diabetes diagnosis. The study was conducted between 2006-2007. Despite the study ending over 10 years ago, steroid use in management has not changed and therefore this study is still relevant. This study recommends conducting CBG monitoring every 4-6 hours after steroid treatment, as their evidence found that testing at a longer interval after steroid treatment meant a patient was more likely to be hyperglycaemic. This is linked to the biological half-life of steroids, which varies with the specific steroid being administered, and may have to be addressed on an individual basis.
Harris et al (2013) also conclude that CBG testing would not lead to increased length of hospital stay as this can be conducted at steroid commencement, and the patient can be provided with instructions for monitoring and timings. This is in agreement with Kwon and Hermayer (2013) who, after conducting a literature review of glucocorticoid-induced hyperglycaemia, concluded that when patients finish antidiabetic medication, their CBG should be monitored as the medication may need to be restarted. Hwangbo and Lee (2017) describe oral glucose loading as the most precise method, but it may not feasible in a clinical setting, is likely to underestimate the prevalence of SIDM which predominately occurs in the evening, and would not assess the immediate effects of recently starting steroids. JBDS (2014) advises testing once a day prior to or following lunch for inpatients. Hwangbo and Lee (2017) advise standardising testing four times a day, which could eliminate transient hyperglycaemia and allow development a trend on which to base management guidelines. This agrees with guidance from JBDS (2014) for the management of outpatient monitoring. Schultz et al (2018) report that in patients who require insulin to manage hyperglycaemia, blood glucose values are high at any time of the day.
In an outpatient department where appointment times can be varied, these high-risk patients would still be identified regardless of CBG timing or frequency monitored. A combination of baseline HbA1c measurement and plasma glucose at the time of a cancer diagnosis, and measuring CBG between one and four times per day would be appropriate and achievable in practice. This would hopefully lead to reduced SIDM admission rates.
Transient steroid-induced diabetes
Hyperglycaemia may be transient, and if it is, do we need to treat it? Within the literature, this question has not been fully answered, as endocrinologists state that there needs to be tight control of CBG, but randomised controlled clinical trials with a large number of patients are lacking (Bonaventura and Montecucoo, 2018).
Jeong et al (2016) explores the incidence of SIDM in patients receiving chemotherapy. This study assessed the incidence of SIDM in 77 patients with no history of diabetes prior to using steroids as an antiemetic, and is one of the few articles which distinguishes between pre-existing diabetes and new onset. The study found that 20% of patients had SIDM after one cycle of treatment, but state this would have been higher if measurements of transient high blood glucose had been included. The study does not identify how many patients may have experienced transient hyperglycaemia or specify a criterion for what is classed as transient, which would have been a useful tool in practice. The JBDS (2014) advise that short periods of hyperglycaemia from short courses of steroids may not require intervention. This raises the question: if blood glucose levels are not monitored, how can it be established if this period of hyperglycaemia is short and therefore transient? A period of hyperglycaemia may revert to normal levels, but may not be fully resolved (Bonaventura and Montecucoo, 2018; JBDS, 2014). The answer is likely to be to monitor blood glucose in patients that are taking short-course steroids, to ensure hyperglycaemia is transient and not a long-term effect (Kwon and Hermayer, 2013).
Impact on service users: risk stratification
Evidence-based guidelines are necessary to ensure that patients are receiving consistent and appropriate care (Storey et al, 2017). Richards et al (2018) describe transparent evidence-based guidance as a pivotal driver for achieving positive outcomes in their assessment of a national approach for improving cancer services. The development of evidence-based guidance will require focus on defining hyperglycaemia and producing a pilot study to assess glucose monitoring in SIDM. This could also lead to the development of risk stratification for outpatients with no previous diabetes diagnosis who are starting treatment with steroids.
The JBSD (2014) advise monitoring high-risk patients, but do not specify what is classed as high risk. Generating a risk stratification for this patient group could allow tailored monitoring and early intervention, but can a consensus be achieved? Rowbottom et al (2015) states that it would be difficult to establish risk factors in this case, as common risk factors for diabetes are not consistent with developing SIDM, which is also reflected in the literature. Hwangbo and Lee (2017) report weight gain, ethnicity, age, sex and disease as factors influencing SIDM. Lee et al (2013) found that old age, obesity and high HbA1c were consistent factors for SIDM. However, Schultz et al (2018) did not find age, BMI or family history to increase risk of developing SIDM. Studies have identified total steroid dose as a risk but do not clearly ascertain guidelines for classing a dose as high (Jeong et al, 2016; Kwon and Hermayer; 2013). Jeong et al. (2016) monitored 8 hour fasting glucose and HbA1c of patients prior to chemotherapy. After 3 and 6 months, they reported that 20% of patients developed SIDM after the first cycle and that a cumulative dose of steroids was statistically significantly associated with deveoloping SIDM. (P=0.049). Tests were performed prior to steroid administration to exclude transient rises in blood glucose. This study identified the likelihood of development of SIDM for the lower dose of ≤156mg and ≥156mg, as 9.1% and 31.8% respectively. In practice, it is common for patients with brain primaries or metastasis to be on daily doses of 8mg once to twice a day, therefore patients would only need to be on steroids for 19 days (≤156mg) or 8.5 days (≥156mg) to be at risk of developing SIDM. In practice, a pilot risk stratification would need to be generated, with cumulative steroid dose and initial CBG readings as risk factors. This could aid in the diagnosis and management of patients that require monitoring, and reduce the chance of over-treatment in patients that do not.
Steroid-induced diabetes awareness
Jeong et al (2016) describes SIDM as a side effect that is often overlooked. Explanations for this could be linked to clinician prioritisation, lack of awareness, or misinterpretation of hyperglycaemic symptoms (Hwangbo et al, 2018). Hyperglycaemia can be asymptomatic until it becomes marked, but symptoms are broad (Jeong et al. 2016). According to the NHS (2018), hyperglycaemia symptoms include increased thirst, urinary frequency, tiredness, blurred vision, weight loss, recurrent infections, abdominal pain, nausea or vomiting and a fruity breath. Many of these symptoms can be attributed to treatment side effects, cancer, co-morbidities or the effects of medication, thus potentially leading to a delay in diagnosis (Jeong et al, 2016). These symptoms have a negative effect on patients and can reduce their quality of life (Gallo et al, 2018). In particular hyperglycaemia can leave patients vulnerable to infection, prolonged hospital stays, and increased mortality (Schultz et al, 2018).
Staff education is paramount to promoting patient education, awareness and compliance. Sampayo and Tofthagen (2017) developed an algorithm to guide nurses through the process from SIDM assessment to intervention, with the aim of improving nursing education. They advocated collaborative working and patient education (Sampayo and Tofthagen, 2017). The core message in the education programme was: glycaemic control and early intervention improve morbidity and mortality outcomes, and are realistic in practice. The NHS Commissioning Board Clinical Reference Group (2013) advocate multidisciplinary care to ensure a seamless locally-provided patient pathway and follow-up; therefore education is a requirement for the whole multidisciplinary team. Financial issues in the NHS is a driver for change within staff models, as traditional roles and responsibilities at all levels will need to be challenged to allow efficient use of resources. The development of radiographer's roles into advanced clinical practioners and consultant practioners is one such change (Society of Radiographers, 2016). This staff group is perfectly situated to affect change in common practice, as these roles are innovative and designed to promote change (Society of Radiographers, 2016). A national shortage of clinical oncology registrars will subsequently lead to a deficit in consultant clinical oncologists; this deficit would lead to more role development, and more advanced practitioners such as consultant radiographers (Royal College of Radiologists, 2015). This provides the perfect opportunity for influence and change in practice. Education of staff in these extended roles would allow for dissemination of education regarding SIDM, which is pertinent for those who are prescribing steroids as independent prescribers, to ensure that these patients have prompt monitoring for SIDM. This dissemination of knowledge, awareness and training, would facilitate incorporation into staff skills, ensuring that staff express a broad skills mix.
The DoH (2018) encourages active patient self-management, and emphasises that much of ill health is preventable. Specific behaviour change techniques are recommended by the DoH (2016), including goal-setting and planning, self-monitoring, relapse prevention, and engaging in social support. Empowering patients make healthier life choices and advocating these behaviour change techniques would help drive change. Personalised patient prevention programmes should be initiated to help people manage their own conditions (DoH, 2018). The objectives set out in the NHS Mandate (DoH, 2018a) would support the change required in practice. The impact on service users would be a reduced number of hyperglycaemic events, which would have a knock-on effect, improving efficiency and productivity through promotion of out-of-hospital care, whilst devising and embracing methods for innovation and growth. Accepting the initial set-up costs of glucose monitoring, there is a financial incentive, as monitoring could lead to reduced in-patient attendances and stay. Despite the potential benefits to service users, negative effects on patients must be considered during this overwhelming time. Staff should encourage and empower patients, leading to improved patient experiences, patient-led decisions about their care and promoting self-management (DoH 2018).
Conclusion
Steroids increase blood glucose levels and cause new hyperglycaemia in patients with no known history of diabetes. The prevalence of SIDM is difficult to establish, but has been found to be statistically significant (Storey et al, 2017; Schultz et al, 2018).
Development of a pilot risk stratification is needed, with cumulative steroid dose and initial CBG readings as potential risks. This could aid in the diagnosis and management of patients where this is required, and reduce the chance of over-treatment in patients that do not require it. This could be combined with altering current CBG tolerances in outpatient protocols, ensuring appropriate identification of patients. Screening patients by testing HbA1c at cancer diagnosis would be appropriate and achievable in practice, and would identify any undiagnosed SIDM. This would lead to reduced SIDM admission rates, which is positive for service users and inpatient targets. Dissemination of knowledge by increasing awareness and staff and patient training would facilitate the incorporation into staff skills, ensuring that staff express a broad skills mix, thus preparing for the NHS ambition of achieving seven-day working week.
Staff should encourage and empower patients, leading to improved patient experience, patient-led decisions about their care and promoting self-management (DoH, 2018). The above areas identified could provide a positive change in current and future practice and is pertinent and attainable in the authors outpatient radiotherapy department.
Key Points
- Blood glucose levels can be increased by steroids, cause new hyperglycaemia in patients with no previous history of diabetes. The prevalence of SIDM is difficult to establish but has been found to be statistically significant
- Development of a pilot risk stratification should be generated, with cumulative steroid dose and initial CBG readings as risks. This could aid in the diagnosis and management of patients that require it, and reduce the chance of over-treatment in patients that do not
- HbA1c screenings at cancer diagnosis would be appropriate and achievable in practice. Doing so would identify any undiagnosed diabetes mellitusStaff are key to encourage and empower patients, leading to improved patient experience, patient-led decisions about their care and promoting self-management (DoH, 2018).
CPD reflective questions
- In what situation would you prescribe steroids?
- What are the national guidelines for steroid prescribing?
- How has this paper influenced your prescribing and/or management of patients?
- What other diagnostic tests could be beneficial?