REVIEW URRENT C OPINION
Glycaemic control in end-of-life care: fundamental or futile? Magnus Lindskog a,b, Lars Ka¨rvestedt c, and Carl Johan Fu¨rst d
Purpose of review Diabetes mellitus is one of the most common comorbidities in palliative care. Yet, the optimal handling of diabetes mellitus in dying patients is debated. This review aims to discuss comprehensively the scientific basis as of today for diabetes mellitus management decisions in end-of-life (EOL) care. Recent findings Glycaemic control provides prognostic information in EOL care of diabetes mellitus patients. Original data on how to manage dying patients with type 2 diabetes mellitus are scarce. Findings in elderly type 2 diabetes mellitus patients and expert opinions support that glycaemic control should be relaxed in dying patients with type 2 diabetes mellitus, in the absence of risk factors for true insulin dependence, to avoid symptomatic hypoglycaemia. For terminal but conscious type 1 diabetes mellitus patients, regular blood glucose measurements and continued insulin therapy is the mainstay, with some discrepancy in preferred management between palliative care physicians and diabetes consultants. No randomized controlled trials are available. Improvement is clearly needed with regard to communication about diabetes mellitus in EOL and documentation of decisions. Corticosteroid-induced diabetes mellitus is a significant problem in palliative care, but predictors exist. Summary In the absence of large observational studies or randomized controlled trials, the current body of knowledge is based on expert opinions, surveys and retrospective studies. Nevertheless, some clinically meaningful recommendations can be made. Prospective studies need to be performed in order to improve our understanding about diabetes mellitus management in EOL. The palliative care community has a joint responsibility to address these questions. Keywords diabetes, end of life, evidence-based, glucose, palliative care
INTRODUCTION Diabetes (diabetes mellitus) is common in the elderly population and associated with cardiovascular disease and cancer and as such one of the most frequent comorbidities encountered in palliative care [1]. A retrospective study recently described the prevalence of diabetes mellitus and the characteristics of diabetes mellitus patients compared with nondiabetes mellitus patients in a palliative care centre. Although one of four patients had a diabetes mellitus diagnosis, the diabetes by itself did not seem to modify prognosis in EOL care. Diabetes mellitus patients had different and more complex care needs [2 ]. The combination of a progressive life-limiting disease and diabetes mellitus raises several questions of metabolic, pharmacological and ethical nature. Although few original studies exist to guide decision &
www.supportiveandpalliativecare.com
making, several expert recommendations have been published. Recently, novel findings have been reported on the basis of retrospective or observational studies, which could have clinical consequences and therefore should be discussed.
a Department of Radiology, Oncology and Radiation Science, Unit of Oncology, Uppsala University, Uppsala, bDepartment of Oncology and Pathology, cDepartment of Molecular Medicine and Surgery, Karolinska Institutet, Solna and dPalliativt utvecklingscentrum, Department of Clinical Sciences, Lund University, Lund, Sweden
Correspondence to Magnus Lindskog, MD PhD, Department of Oncology, Akademiska University Hospital, Building 78, Ground floor, 751 85 Uppsala, Sweden. Tel: +46 73 349 0069; e-mail: magnus.lindskog@ ki.se Curr Opin Support Palliat Care 2014, 8:378–382 DOI:10.1097/SPC.0000000000000095 Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Glycaemic control in end-of-life care Lindskog et al.
KEY POINTS Original data on diabetes management in end of life are very limited, warranting novel well designed prospective studies. Hypoglycaemia is likely the main metabolic threat to dying patients with T2DM. Insulin should generally be continued in T1DM diabetes as long as the patient is conscious; the optimal insulin schedule remains to be established. Communication between different healthcare professionals caring for dying patients with diabetes needs to be improved and so does the communication with the patients themselves and their families.
The purpose of this review was to comprehensively evaluate the scientific basis for existing recommendations from a palliative care perspective, integrating the most current data about diabetes mellitus management in dying patients. Relevant publications were identified by searching PubMed for articles in English, from 1969 to 2014, with emphasis on publications from the last 10 years. Search terms included the keywords diabetes, antidiabetic therapy, hyperglycaemia or hypoglycaemia, oral hypoglycaemic agent, insulin, in combinations with one or several of the keywords palliative, hospice, end of life, dying, terminal illness.
BLOOD GLUCOSE MONITORING IN END-OF-LIFE CARE Most authors agree that tight control of blood glucose is neither important nor desirable in terminally ill patients as long as the glucose level is in the range of 5–20 mml/l [3–9]. Nevertheless, a retrospective study of patients with diabetes mellitus who died of cancer showed that monitoring of blood glucose was continued until death in 76% of the cases. Nineteen percent of the study population had type 1 diabetes mellitus (T1DM) [5]. In a review on management of diabetes mellitus in patients with advanced cancer, Poulson suggested an algorithm for glucose monitoring, taking into consideration diabetes mellitus subtype, nutritional status and capacity for oral intake, and whether the patient was terminal or not. Although easy to use, the scheme was based on personal experience [4]. Other expert opinions have been published, most of which favour a similar approach, however none of them is evidence-based.
PROFESSIONAL EXPERIENCES ACROSS THE HEALTHCARE SPECTRUM In their pivotal study from 2006, Ford-Dunn et al. [6] challenged a large number of consultant
diabetologists (DC) and palliative care consultants (PCC), respectively, in the United Kingdom with a questionnaire based around three case scenarios of imminently dying diabetes mellitus patients. In the scenario dealing with a dying patient with type 2 diabetes mellitus (T2DM) controlled by oral antidiabetic drugs (OAD), about 80% of PCC and 70% of DC chose to stop both blood glucose monitoring and treatment. When the scenario was changed to a dying T2DM patient whose diabetes was controlled by intermediate mixed insulin, a majority of both groups still advised to both stop treatment and blood glucose monitoring. Every third DC and four out of 10 PCC would, however, continue insulin therapy. In such a situation, most PCC favoured switching to short-acting insulin as needed (PRN) whereas DC generally favoured continuing intermediate mixed insulin. In the third and last scenario, describing a patient with insulin-controlled T1DM, 25% of DC and 30% of PCC would stop therapy and blood glucose monitoring. Again, DC were much more prone to continue intermediate-acting insulin whereas PCC favoured short-acting insulin PRN [6]. Whereas the more relaxed attitudes of PCC, favouring the use of PRN insulin injections only, may seem more adapted to the nature of EOL care, the argument of DC that one insulin dose with flat profile more accurately mimics normal physiology in the fasting state and, therefore, minimizes the risk of hypoglycaemia is worth considering. A once daily injection could be seen as an advantage from ethical (less often painful finger pricking) and logistic (less frequent instrumentation and documentation) points of view. No comparative studies between short-acting and long-acting insulin in dying patients with T1DM have been conducted. Quinn et al. [7] undertook focus groups’ discussions and a cross-sectional survey to explore the professional experiences of doctors and nurses caring for dying diabetes mellitus patients. The survey included endocrinologists, palliative care physicians, general practitioners and nurses. Achieving comfort, whilst limiting the need for intrusive investigations, was identified as the goal of care as well as the most important challenge. Fifty-four percent reported that they used an individualized approach, integrating information about the patient’s medical situation, nutritional and mental state, blood glucose, the patient’s wishes and views of the family in the management of diabetes mellitus in the last week of life. Ninety percent of participants regarded blood glucose monitoring as important in EOL care. During the last days of life, a majority chose to monitor blood glucose if the patient had symptoms suspicious of hyperglycaemia. Thirteen percent
1751-4258 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.supportiveandpalliativecare.com
379
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
End of life management
continued daily blood glucose measurements until the patient died. The lack of guidelines was found to be problematic, in particular by doctors and nurses in palliative care [7].
HBA1C Although HbA1C has generally been view as irrelevant in palliative care [8,10,11], Kondo et al. recently identified metabolic control defined as the A1C value above or below 7.5% at the time of admission to hospice, as a prognostic factor in advanced cancer patients with preexisting T2DM. The median length of the EOL phase was significantly longer in diabetes mellitus patients with good metabolic control. The prognostic effect was independent of performance status [12 ]. &&
COMMUNICATING DIFFERENT OPTIONS The ethics of finger pricking is a common theme in surveys of doctors and nurses who treat diabetes mellitus patients in EOL. Some argue that finger pricking is a futile, invasive and painful procedure that distresses the patient. Others disagree, arguing that compared with subcutaneous injections of drugs, urine catheters, etc. that are frequently used in EOL care, finger pricking may not be more invasive or distressing and should be seen as a screening instrument for treatable symptoms [5,7,9]. Moreover, patients and families are typically accustomed to daily blood glucose measurements and may regard the cessation of glycaemic monitoring as a sign of poor standard of care or an indication of the palliative care staff ‘giving up’ on their loved one. Therefore, communicating with the family early on, explaining the rational for stopping routine testing of glucose when that is considered appropriate and emphasizing the shift in focus of the care to ensure maximum comfort for the patient are all important. One study found medical decisions about antidiabetic therapy, as well as communication with patients or families about the diabetes mellitus management, to be poorly documented in the charts [5]. A recent review by Dikkers et al. identified five key themes about the information needs of family carers of diabetes mellitus patients at the end of life. Typically, the family needs to be approached about how to perform diabetes care tasks, be informed about the shift in the focus and goals of care when approaching death, be involved in discussions about principles of blood glucose management, be educated about what can be expected as death approaches and be actively involved in decisions about diabetes care [13 ]. A recent study reported that glucose monitoring in EOL did not seem overly &&
380
www.supportiveandpalliativecare.com
burdensome to most patients, albeit the monitoring was stopped when death was imminent [14].
HYPOGLYCAEMIA AS THE MAIN RISK FOR ELDERLY AND DYING T2DM PATIENTS Original studies on the relevance of glycaemic control for the well being of diabetes mellitus patients receiving EOL care are lacking. Data from elderly diabetes mellitus patients in nonpalliative care settings suggest that they are often overtreated, with frequent hypoglycaemias despite regular eating habits and good compliance with medication [15,16]. Hypoglycaemia is typically associated with several unpleasant symptoms including sweating, anxiety, tremors, weakness, palpitations, and can cause seizures or even be fatal. Furthermore, manifest hypoglycaemia may not easily be corrected in terminally ill patients who have lost the oral route, may lack robust venous access and have empty glycogen stores [4,17]. Notably, hyperglycaemia may cause highly uncomfortable symptoms as a result of dehydration including restlessness, nausea, weakness, blurred vision, confusion or even coma, as well as metabolic acidosis [18]. In T2DM patients with progressive life-limiting disease, about two-thirds are prescribed OAD whereas one third requires insulin [5]. One study found that elderly patients with tightly controlled T2DM were unlikely to experience any significant elevation of blood glucose during the first month after discontinuation of OAD. At the same time, the risk of hypoglycaemic events was diminished [19]. In patients with poorly controlled T2DM, withdrawal of OAD may cause rapid metabolic deterioration. Nonketotic hyperosmolar syndrome is the single most relevant complication of loss of metabolic control in T2DM. However, 32% of all episodes of ketoacidosis occur in patients with T2DM, although usually less severe than in T1DM [20]. As end of life approaches, renal function declines, impairing the clearance of insulin and long-acting sulphonylureas that predisposes hypoglycaemia. Metformin often causes nausea and vomiting and has a risk of fatal lactic acidosis in the event of drug accumulation due to kidney failure. Thus, common antidiabetic drugs may cause or aggravate severe symptoms in EOL. Angelo et al. [8] recently suggested an algorithm for consideration of different OAD as well as insulin in a palliative care setting. In a recent international survey on experts’ opinions regarding potentially inappropriate medications for dying patients, 81% considered OAD to be inappropriate [21 ]. Taken together, most reports favour a view of hypoglycaemias as being more likely to &
Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Glycaemic control in end-of-life care Lindskog et al.
occur and more dangerous to the patient at end of life, than episodes of symptomatic hyperglycaemia, the latter possible to rapidly control by insulin PRN. Therefore, despite the multitude of available and efficient OAD, the principle of do-no-harm together with the concept of neither shortening nor prolonging life, inherent to palliative care philosophy, urge us to stay away from OAD when caring for a dying patient with T2DM. In a recent study, Vandenhaute [22] demonstrates that the international recommendations for T2DM management, though scientifically based, are hardly relevant at all in palliative care and could even compromise the very principles that should govern EOL care.
CORTICOSTEROID-INDUCED DIABETES IN END-OF-LIFE CARE Among iatrogenic causes of hyperglycaemia in palliative care, corticosteroids deserve special attention. In terms of achieving control of a multitude of the symptoms that are common in advanced disease, corticosteroids are outstanding, which explains why corticosteroids are prescribed to as many as 30–60% of all patients receiving palliative care. The induction of diabetes mellitus by corticosteroids is caused by multiple factors. A recent retrospective study by Pilkey et al. found corticosteroid-induced diabetes mellitus in palliative care to be more common than previously thought. Looking specifically at dexamethasone use, a dosedependent increase in diabetes mellitus risk was found; although patients treated with lower corticosteroid doses were also at risk. In contrast, no correlation was found with duration of corticosteroid treatment. Corticosteroid-induced diabetes mellitus was significantly more common in patients with brain tumours compared with other fatal diagnoses. Almost all patients remained on corticosteroids at the time of their deaths [14].
CONTINUED INSULIN TO DYING DIABETICS: INEVITABLE OR INACCURATE? Although randomized data are lacking, published literature uniformly advocates that once a dying patient becomes unconscious measurements of blood glucose and insulin injections should be permanently discontinued irrespective of diabetes mellitus type, provided that the loss of consciousness cannot be explained by an extreme alteration in blood glucose [6]. If in any doubt, blood sugar could be checked. The family of the dying patient needs to be carefully informed because insulin is usually regarded as life sustaining. Secondly, the diabetes
mellitus subtype should be considered before deciding whether and how to proceed with insulin therapy in conscious but imminently dying patients. The basis for this distinction is that T1DM patients need to receive insulin daily to prevent the development of ketoacidosis also if the patient stops eating. In EOL care, most palliative care physicians would argue that the presence or absence of persisting symptoms that can be linked to hyperglycaemia is the real issue, not plasma glucose level, dehydration or acidosis per se. In obese patients with T2DM, fasting and postprandial blood glucose levels were not significantly affected during 2 weeks following cessation of insulin therapy [23]. The issue could be more complicated if the patient in fact has a truly insulin-dependent T2DM. Boyd [3] suggested that one or more of the following would indicate the continued need for insulin in T2DM: first, introduction of insulin soon after the diagnosis of diabetes; second, a history of previous ketoacidosis; or third, the use of more than one daily dose of insulin for many years. For patients with T1DM in EOL, virtually no original data exists. In order to maintain euglycaemia and to avoid severely symptomatic acidosis or even hastened death, most authors favour regularly measurements of blood glucose in T1DM patients [5,6,8]. Different views on optimal insulin schedules exist, as evident from the paper of Ford-Dunn discussed earlier. Whatever schedule, the insulin dose needs to be reduced as the patient looses weight or has a reduced calorie intake due to anorexia or vomiting. No comparative studies between shortacting and long-acting insulin have been conducted in dying T1DM patients.
CONCLUSION Undoubtedly, very few studies of high scientific standards have addressed how diabetes mellitus management in EOL care could be optimised. The palliative care community has a responsibility to conduct prospective studies to improve our understanding of terminal care for this large group of patients. Research questions include, for example, to investigate how diabetes mellitus patients with advanced life-limiting disease feel themselves about finger pricking and insulin therapy as they progress to terminal care, that is patient-reported experience measures. For insulin-dependent T2DM receiving palliative care, a well powered prospective randomized trial investigating if continued insulin therapy is superior or not to insulin PRN would be highly informative. Data on symptoms and patient satisfaction should be systematically collected in addition to blood glucose monitoring. Finally, in
1751-4258 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.supportiveandpalliativecare.com
381
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
End of life management
T1DM, different insulin schedules could be evaluated prospectively with respect to blood glucose control and symptom development in patients receiving palliative care who are not imminently dying. The time has come to evidence-based end-of-life care for diabetes patients. What are we waiting for? Acknowledgements None. Conflicts of interest The authors have no conflict of interest to declare.
REFERENCES AND RECOMMENDED READING Papers of particular interest, published within the annual period of review, have been highlighted as: & of special interest && of outstanding interest 1. Dybicz SB, Thompson S, Molotsky S, Stuart B. Prevalence of diabetes and the burden of comorbid conditions among elderly nursing home residents. Am J Geriatr Pharmacother 2011; 9:212–223. 2. Dionisio R, Giardini A, Cata PD, et al. Diabetes management in end of life: a & preliminary report stemming from clinical experience. Am J Hosp Palliat Care 2014. [Epub ahead of print] A retrospective study that recently described the prevalence of diabetes mellitus and the characteristics of diabetes mellitus patients compared with non-diabetes mellitus patients in a palliative care centre. 3. Boyd K. Diabetes mellitus in hospice patients: some guidelines. Palliat Med 1993; 7:163–164. 4. Poulson J. The management of diabetes in patients with advanced cancer. J Pain Symptom Manage 1997; 13:339–346. 5. McCoubrie R, Jeffrey D, Paton C, Dawes L. Managing diabetes mellitus in patients with advanced cancer: a case note audit and guidelines. Eur J Cancer Care (Engl) 2005; 14:244–248. 6. Ford-Dunn S, Smith A, Quin J. Management of diabetes during the last days of life: attitudes of consultant diabetologists and consultant palliative care physicians in the UK. Palliat Med 2006; 20:197–203. 7. Quinn K, Hudson P, Dunning T. Diabetes management in patients receiving palliative care. J Pain Symptom Manage 2006; 32:275–286.
382
www.supportiveandpalliativecare.com
8. Angelo M, Ruchalski C, Sproge BJ. An approach to diabetes mellitus in hospice and palliative medicine. J Palliat Med 2011; 14:83–87. 9. King EJ, Haboubi H, Evans D, et al. The management of diabetes in terminal illness related to cancer. QJM 2012; 105:3–9. 10. McPherson ML. Management of diabetes at end of life. Home Healthc Nurse 2008; 26:276–278. 11. Jeffreys E, Rosielle DA. Diabetes management at the end of life #258. J Palliat Med 2012; 15:1152–1154. 12. Kondo S, Kondo M, Kondo A. Glycemia control using A1C level in terminal && cancer patients with preexisting type 2 diabetes. J Palliat Med 2013; 16:790–793. This study recently identified metabolic control defined as the A1C value above or below 7.5% at the time of admission to hospice, as a prognostic factor in advanced cancer patients with preexisting T2DM. 13. Dikkers MF, Dunning T, Savage S. Information needs of family carers of && people with diabetes at the end of life: a literature review. J Palliat Med 2013; 16:1617–1623. A recent review that identified five key themes about the information needs of family carers of diabetes mellitus patients at the end of life. 14. Pilkey J, Streeter L, Beel A, et al. Corticosteroid-induced diabetes in palliative care. J Palliat Med 2012; 15:681–689. 15. Lo¨fgren UB, Rosenqvist U, Lindstro¨m T, et al. Diabetes control in Swedish community dwelling elderly: more often tight than poor. J Intern Med 2004; 255:96–101. 16. Bouillet B, Vaillant G, Petit JM, et al. Are elderly patients with diabetes being overtreated in French long-term-care homes? Diabetes Metab 2010; 36:272–277. 17. Taborsky GJ Jr. The physiology of glucagon. J Diabetes Sci Technol 2010; 4:1338–1344. 18. Warren RE, Deary IJ, Frier BM. The symptoms of hyperglycaemia in people with insulin-treated diabetes: classification using principal components analysis. Diabetes Metab Res Rev 2003; 19:408–414. 19. Sjo¨blom P, Tengblad A, Lo¨fgren UB, et al. Can diabetes medication be reduced in elderly patients? An observational study of diabetes drug withdrawal in nursing home patients with tight glycaemic control. Diabetes Res Clin Pract 2008; 82:197–202. 20. Wang ZH, Kihl-Selstam E, Eriksson JW. Ketoacidosis occurs in both Type 1 and Type 2 diabetes–a population-based study from Northern Sweden. Diabet Med 2008; 25:867–870. 21. Raijmakers NJ, van Zuylen L, Furst CJ, et al. Variation in medication use in & cancer patients at the end of life: a cross-sectional analysis. Support Care Cancer 2013; 21:1003–1011. As pointed out by Raijmakers et al. in a recent international survey on potentially inappropriate medications for dying patients, 81% of experts considered OAD to be inappropriate. 22. Vandenhaute V. Palliative care and type II diabetes: a need for new guidelines? Am J Hosp Palliat Care 2010; 27:444–445. 23. Andrews WJ, Vasquez B, Nagulesparan M, et al. Insulin therapy in obese, noninsulin-dependent diabetes induces improvements in insulin action and secretion that are maintained for two weeks after insulin withdrawal. Diabetes 1984; 33:634–642.
Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Glycaemic control in end-of-life care: fundamental or futile? Magnus Lindskog a,b, Lars Ka¨rvestedt c, and Carl Johan Fu¨rst d
Purpose of review Diabetes mellitus is one of the most common comorbidities in palliative care. Yet, the optimal handling of diabetes mellitus in dying patients is debated. This review aims to discuss comprehensively the scientific basis as of today for diabetes mellitus management decisions in end-of-life (EOL) care. Recent findings Glycaemic control provides prognostic information in EOL care of diabetes mellitus patients. Original data on how to manage dying patients with type 2 diabetes mellitus are scarce. Findings in elderly type 2 diabetes mellitus patients and expert opinions support that glycaemic control should be relaxed in dying patients with type 2 diabetes mellitus, in the absence of risk factors for true insulin dependence, to avoid symptomatic hypoglycaemia. For terminal but conscious type 1 diabetes mellitus patients, regular blood glucose measurements and continued insulin therapy is the mainstay, with some discrepancy in preferred management between palliative care physicians and diabetes consultants. No randomized controlled trials are available. Improvement is clearly needed with regard to communication about diabetes mellitus in EOL and documentation of decisions. Corticosteroid-induced diabetes mellitus is a significant problem in palliative care, but predictors exist. Summary In the absence of large observational studies or randomized controlled trials, the current body of knowledge is based on expert opinions, surveys and retrospective studies. Nevertheless, some clinically meaningful recommendations can be made. Prospective studies need to be performed in order to improve our understanding about diabetes mellitus management in EOL. The palliative care community has a joint responsibility to address these questions. Keywords diabetes, end of life, evidence-based, glucose, palliative care
INTRODUCTION Diabetes (diabetes mellitus) is common in the elderly population and associated with cardiovascular disease and cancer and as such one of the most frequent comorbidities encountered in palliative care [1]. A retrospective study recently described the prevalence of diabetes mellitus and the characteristics of diabetes mellitus patients compared with nondiabetes mellitus patients in a palliative care centre. Although one of four patients had a diabetes mellitus diagnosis, the diabetes by itself did not seem to modify prognosis in EOL care. Diabetes mellitus patients had different and more complex care needs [2 ]. The combination of a progressive life-limiting disease and diabetes mellitus raises several questions of metabolic, pharmacological and ethical nature. Although few original studies exist to guide decision &
www.supportiveandpalliativecare.com
making, several expert recommendations have been published. Recently, novel findings have been reported on the basis of retrospective or observational studies, which could have clinical consequences and therefore should be discussed.
a Department of Radiology, Oncology and Radiation Science, Unit of Oncology, Uppsala University, Uppsala, bDepartment of Oncology and Pathology, cDepartment of Molecular Medicine and Surgery, Karolinska Institutet, Solna and dPalliativt utvecklingscentrum, Department of Clinical Sciences, Lund University, Lund, Sweden
Correspondence to Magnus Lindskog, MD PhD, Department of Oncology, Akademiska University Hospital, Building 78, Ground floor, 751 85 Uppsala, Sweden. Tel: +46 73 349 0069; e-mail: magnus.lindskog@ ki.se Curr Opin Support Palliat Care 2014, 8:378–382 DOI:10.1097/SPC.0000000000000095 Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Glycaemic control in end-of-life care Lindskog et al.
KEY POINTS Original data on diabetes management in end of life are very limited, warranting novel well designed prospective studies. Hypoglycaemia is likely the main metabolic threat to dying patients with T2DM. Insulin should generally be continued in T1DM diabetes as long as the patient is conscious; the optimal insulin schedule remains to be established. Communication between different healthcare professionals caring for dying patients with diabetes needs to be improved and so does the communication with the patients themselves and their families.
The purpose of this review was to comprehensively evaluate the scientific basis for existing recommendations from a palliative care perspective, integrating the most current data about diabetes mellitus management in dying patients. Relevant publications were identified by searching PubMed for articles in English, from 1969 to 2014, with emphasis on publications from the last 10 years. Search terms included the keywords diabetes, antidiabetic therapy, hyperglycaemia or hypoglycaemia, oral hypoglycaemic agent, insulin, in combinations with one or several of the keywords palliative, hospice, end of life, dying, terminal illness.
BLOOD GLUCOSE MONITORING IN END-OF-LIFE CARE Most authors agree that tight control of blood glucose is neither important nor desirable in terminally ill patients as long as the glucose level is in the range of 5–20 mml/l [3–9]. Nevertheless, a retrospective study of patients with diabetes mellitus who died of cancer showed that monitoring of blood glucose was continued until death in 76% of the cases. Nineteen percent of the study population had type 1 diabetes mellitus (T1DM) [5]. In a review on management of diabetes mellitus in patients with advanced cancer, Poulson suggested an algorithm for glucose monitoring, taking into consideration diabetes mellitus subtype, nutritional status and capacity for oral intake, and whether the patient was terminal or not. Although easy to use, the scheme was based on personal experience [4]. Other expert opinions have been published, most of which favour a similar approach, however none of them is evidence-based.
PROFESSIONAL EXPERIENCES ACROSS THE HEALTHCARE SPECTRUM In their pivotal study from 2006, Ford-Dunn et al. [6] challenged a large number of consultant
diabetologists (DC) and palliative care consultants (PCC), respectively, in the United Kingdom with a questionnaire based around three case scenarios of imminently dying diabetes mellitus patients. In the scenario dealing with a dying patient with type 2 diabetes mellitus (T2DM) controlled by oral antidiabetic drugs (OAD), about 80% of PCC and 70% of DC chose to stop both blood glucose monitoring and treatment. When the scenario was changed to a dying T2DM patient whose diabetes was controlled by intermediate mixed insulin, a majority of both groups still advised to both stop treatment and blood glucose monitoring. Every third DC and four out of 10 PCC would, however, continue insulin therapy. In such a situation, most PCC favoured switching to short-acting insulin as needed (PRN) whereas DC generally favoured continuing intermediate mixed insulin. In the third and last scenario, describing a patient with insulin-controlled T1DM, 25% of DC and 30% of PCC would stop therapy and blood glucose monitoring. Again, DC were much more prone to continue intermediate-acting insulin whereas PCC favoured short-acting insulin PRN [6]. Whereas the more relaxed attitudes of PCC, favouring the use of PRN insulin injections only, may seem more adapted to the nature of EOL care, the argument of DC that one insulin dose with flat profile more accurately mimics normal physiology in the fasting state and, therefore, minimizes the risk of hypoglycaemia is worth considering. A once daily injection could be seen as an advantage from ethical (less often painful finger pricking) and logistic (less frequent instrumentation and documentation) points of view. No comparative studies between short-acting and long-acting insulin in dying patients with T1DM have been conducted. Quinn et al. [7] undertook focus groups’ discussions and a cross-sectional survey to explore the professional experiences of doctors and nurses caring for dying diabetes mellitus patients. The survey included endocrinologists, palliative care physicians, general practitioners and nurses. Achieving comfort, whilst limiting the need for intrusive investigations, was identified as the goal of care as well as the most important challenge. Fifty-four percent reported that they used an individualized approach, integrating information about the patient’s medical situation, nutritional and mental state, blood glucose, the patient’s wishes and views of the family in the management of diabetes mellitus in the last week of life. Ninety percent of participants regarded blood glucose monitoring as important in EOL care. During the last days of life, a majority chose to monitor blood glucose if the patient had symptoms suspicious of hyperglycaemia. Thirteen percent
1751-4258 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.supportiveandpalliativecare.com
379
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
End of life management
continued daily blood glucose measurements until the patient died. The lack of guidelines was found to be problematic, in particular by doctors and nurses in palliative care [7].
HBA1C Although HbA1C has generally been view as irrelevant in palliative care [8,10,11], Kondo et al. recently identified metabolic control defined as the A1C value above or below 7.5% at the time of admission to hospice, as a prognostic factor in advanced cancer patients with preexisting T2DM. The median length of the EOL phase was significantly longer in diabetes mellitus patients with good metabolic control. The prognostic effect was independent of performance status [12 ]. &&
COMMUNICATING DIFFERENT OPTIONS The ethics of finger pricking is a common theme in surveys of doctors and nurses who treat diabetes mellitus patients in EOL. Some argue that finger pricking is a futile, invasive and painful procedure that distresses the patient. Others disagree, arguing that compared with subcutaneous injections of drugs, urine catheters, etc. that are frequently used in EOL care, finger pricking may not be more invasive or distressing and should be seen as a screening instrument for treatable symptoms [5,7,9]. Moreover, patients and families are typically accustomed to daily blood glucose measurements and may regard the cessation of glycaemic monitoring as a sign of poor standard of care or an indication of the palliative care staff ‘giving up’ on their loved one. Therefore, communicating with the family early on, explaining the rational for stopping routine testing of glucose when that is considered appropriate and emphasizing the shift in focus of the care to ensure maximum comfort for the patient are all important. One study found medical decisions about antidiabetic therapy, as well as communication with patients or families about the diabetes mellitus management, to be poorly documented in the charts [5]. A recent review by Dikkers et al. identified five key themes about the information needs of family carers of diabetes mellitus patients at the end of life. Typically, the family needs to be approached about how to perform diabetes care tasks, be informed about the shift in the focus and goals of care when approaching death, be involved in discussions about principles of blood glucose management, be educated about what can be expected as death approaches and be actively involved in decisions about diabetes care [13 ]. A recent study reported that glucose monitoring in EOL did not seem overly &&
380
www.supportiveandpalliativecare.com
burdensome to most patients, albeit the monitoring was stopped when death was imminent [14].
HYPOGLYCAEMIA AS THE MAIN RISK FOR ELDERLY AND DYING T2DM PATIENTS Original studies on the relevance of glycaemic control for the well being of diabetes mellitus patients receiving EOL care are lacking. Data from elderly diabetes mellitus patients in nonpalliative care settings suggest that they are often overtreated, with frequent hypoglycaemias despite regular eating habits and good compliance with medication [15,16]. Hypoglycaemia is typically associated with several unpleasant symptoms including sweating, anxiety, tremors, weakness, palpitations, and can cause seizures or even be fatal. Furthermore, manifest hypoglycaemia may not easily be corrected in terminally ill patients who have lost the oral route, may lack robust venous access and have empty glycogen stores [4,17]. Notably, hyperglycaemia may cause highly uncomfortable symptoms as a result of dehydration including restlessness, nausea, weakness, blurred vision, confusion or even coma, as well as metabolic acidosis [18]. In T2DM patients with progressive life-limiting disease, about two-thirds are prescribed OAD whereas one third requires insulin [5]. One study found that elderly patients with tightly controlled T2DM were unlikely to experience any significant elevation of blood glucose during the first month after discontinuation of OAD. At the same time, the risk of hypoglycaemic events was diminished [19]. In patients with poorly controlled T2DM, withdrawal of OAD may cause rapid metabolic deterioration. Nonketotic hyperosmolar syndrome is the single most relevant complication of loss of metabolic control in T2DM. However, 32% of all episodes of ketoacidosis occur in patients with T2DM, although usually less severe than in T1DM [20]. As end of life approaches, renal function declines, impairing the clearance of insulin and long-acting sulphonylureas that predisposes hypoglycaemia. Metformin often causes nausea and vomiting and has a risk of fatal lactic acidosis in the event of drug accumulation due to kidney failure. Thus, common antidiabetic drugs may cause or aggravate severe symptoms in EOL. Angelo et al. [8] recently suggested an algorithm for consideration of different OAD as well as insulin in a palliative care setting. In a recent international survey on experts’ opinions regarding potentially inappropriate medications for dying patients, 81% considered OAD to be inappropriate [21 ]. Taken together, most reports favour a view of hypoglycaemias as being more likely to &
Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Glycaemic control in end-of-life care Lindskog et al.
occur and more dangerous to the patient at end of life, than episodes of symptomatic hyperglycaemia, the latter possible to rapidly control by insulin PRN. Therefore, despite the multitude of available and efficient OAD, the principle of do-no-harm together with the concept of neither shortening nor prolonging life, inherent to palliative care philosophy, urge us to stay away from OAD when caring for a dying patient with T2DM. In a recent study, Vandenhaute [22] demonstrates that the international recommendations for T2DM management, though scientifically based, are hardly relevant at all in palliative care and could even compromise the very principles that should govern EOL care.
CORTICOSTEROID-INDUCED DIABETES IN END-OF-LIFE CARE Among iatrogenic causes of hyperglycaemia in palliative care, corticosteroids deserve special attention. In terms of achieving control of a multitude of the symptoms that are common in advanced disease, corticosteroids are outstanding, which explains why corticosteroids are prescribed to as many as 30–60% of all patients receiving palliative care. The induction of diabetes mellitus by corticosteroids is caused by multiple factors. A recent retrospective study by Pilkey et al. found corticosteroid-induced diabetes mellitus in palliative care to be more common than previously thought. Looking specifically at dexamethasone use, a dosedependent increase in diabetes mellitus risk was found; although patients treated with lower corticosteroid doses were also at risk. In contrast, no correlation was found with duration of corticosteroid treatment. Corticosteroid-induced diabetes mellitus was significantly more common in patients with brain tumours compared with other fatal diagnoses. Almost all patients remained on corticosteroids at the time of their deaths [14].
CONTINUED INSULIN TO DYING DIABETICS: INEVITABLE OR INACCURATE? Although randomized data are lacking, published literature uniformly advocates that once a dying patient becomes unconscious measurements of blood glucose and insulin injections should be permanently discontinued irrespective of diabetes mellitus type, provided that the loss of consciousness cannot be explained by an extreme alteration in blood glucose [6]. If in any doubt, blood sugar could be checked. The family of the dying patient needs to be carefully informed because insulin is usually regarded as life sustaining. Secondly, the diabetes
mellitus subtype should be considered before deciding whether and how to proceed with insulin therapy in conscious but imminently dying patients. The basis for this distinction is that T1DM patients need to receive insulin daily to prevent the development of ketoacidosis also if the patient stops eating. In EOL care, most palliative care physicians would argue that the presence or absence of persisting symptoms that can be linked to hyperglycaemia is the real issue, not plasma glucose level, dehydration or acidosis per se. In obese patients with T2DM, fasting and postprandial blood glucose levels were not significantly affected during 2 weeks following cessation of insulin therapy [23]. The issue could be more complicated if the patient in fact has a truly insulin-dependent T2DM. Boyd [3] suggested that one or more of the following would indicate the continued need for insulin in T2DM: first, introduction of insulin soon after the diagnosis of diabetes; second, a history of previous ketoacidosis; or third, the use of more than one daily dose of insulin for many years. For patients with T1DM in EOL, virtually no original data exists. In order to maintain euglycaemia and to avoid severely symptomatic acidosis or even hastened death, most authors favour regularly measurements of blood glucose in T1DM patients [5,6,8]. Different views on optimal insulin schedules exist, as evident from the paper of Ford-Dunn discussed earlier. Whatever schedule, the insulin dose needs to be reduced as the patient looses weight or has a reduced calorie intake due to anorexia or vomiting. No comparative studies between shortacting and long-acting insulin have been conducted in dying T1DM patients.
CONCLUSION Undoubtedly, very few studies of high scientific standards have addressed how diabetes mellitus management in EOL care could be optimised. The palliative care community has a responsibility to conduct prospective studies to improve our understanding of terminal care for this large group of patients. Research questions include, for example, to investigate how diabetes mellitus patients with advanced life-limiting disease feel themselves about finger pricking and insulin therapy as they progress to terminal care, that is patient-reported experience measures. For insulin-dependent T2DM receiving palliative care, a well powered prospective randomized trial investigating if continued insulin therapy is superior or not to insulin PRN would be highly informative. Data on symptoms and patient satisfaction should be systematically collected in addition to blood glucose monitoring. Finally, in
1751-4258 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.supportiveandpalliativecare.com
381
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
End of life management
T1DM, different insulin schedules could be evaluated prospectively with respect to blood glucose control and symptom development in patients receiving palliative care who are not imminently dying. The time has come to evidence-based end-of-life care for diabetes patients. What are we waiting for? Acknowledgements None. Conflicts of interest The authors have no conflict of interest to declare.
REFERENCES AND RECOMMENDED READING Papers of particular interest, published within the annual period of review, have been highlighted as: & of special interest && of outstanding interest 1. Dybicz SB, Thompson S, Molotsky S, Stuart B. Prevalence of diabetes and the burden of comorbid conditions among elderly nursing home residents. Am J Geriatr Pharmacother 2011; 9:212–223. 2. Dionisio R, Giardini A, Cata PD, et al. Diabetes management in end of life: a & preliminary report stemming from clinical experience. Am J Hosp Palliat Care 2014. [Epub ahead of print] A retrospective study that recently described the prevalence of diabetes mellitus and the characteristics of diabetes mellitus patients compared with non-diabetes mellitus patients in a palliative care centre. 3. Boyd K. Diabetes mellitus in hospice patients: some guidelines. Palliat Med 1993; 7:163–164. 4. Poulson J. The management of diabetes in patients with advanced cancer. J Pain Symptom Manage 1997; 13:339–346. 5. McCoubrie R, Jeffrey D, Paton C, Dawes L. Managing diabetes mellitus in patients with advanced cancer: a case note audit and guidelines. Eur J Cancer Care (Engl) 2005; 14:244–248. 6. Ford-Dunn S, Smith A, Quin J. Management of diabetes during the last days of life: attitudes of consultant diabetologists and consultant palliative care physicians in the UK. Palliat Med 2006; 20:197–203. 7. Quinn K, Hudson P, Dunning T. Diabetes management in patients receiving palliative care. J Pain Symptom Manage 2006; 32:275–286.
382
www.supportiveandpalliativecare.com
8. Angelo M, Ruchalski C, Sproge BJ. An approach to diabetes mellitus in hospice and palliative medicine. J Palliat Med 2011; 14:83–87. 9. King EJ, Haboubi H, Evans D, et al. The management of diabetes in terminal illness related to cancer. QJM 2012; 105:3–9. 10. McPherson ML. Management of diabetes at end of life. Home Healthc Nurse 2008; 26:276–278. 11. Jeffreys E, Rosielle DA. Diabetes management at the end of life #258. J Palliat Med 2012; 15:1152–1154. 12. Kondo S, Kondo M, Kondo A. Glycemia control using A1C level in terminal && cancer patients with preexisting type 2 diabetes. J Palliat Med 2013; 16:790–793. This study recently identified metabolic control defined as the A1C value above or below 7.5% at the time of admission to hospice, as a prognostic factor in advanced cancer patients with preexisting T2DM. 13. Dikkers MF, Dunning T, Savage S. Information needs of family carers of && people with diabetes at the end of life: a literature review. J Palliat Med 2013; 16:1617–1623. A recent review that identified five key themes about the information needs of family carers of diabetes mellitus patients at the end of life. 14. Pilkey J, Streeter L, Beel A, et al. Corticosteroid-induced diabetes in palliative care. J Palliat Med 2012; 15:681–689. 15. Lo¨fgren UB, Rosenqvist U, Lindstro¨m T, et al. Diabetes control in Swedish community dwelling elderly: more often tight than poor. J Intern Med 2004; 255:96–101. 16. Bouillet B, Vaillant G, Petit JM, et al. Are elderly patients with diabetes being overtreated in French long-term-care homes? Diabetes Metab 2010; 36:272–277. 17. Taborsky GJ Jr. The physiology of glucagon. J Diabetes Sci Technol 2010; 4:1338–1344. 18. Warren RE, Deary IJ, Frier BM. The symptoms of hyperglycaemia in people with insulin-treated diabetes: classification using principal components analysis. Diabetes Metab Res Rev 2003; 19:408–414. 19. Sjo¨blom P, Tengblad A, Lo¨fgren UB, et al. Can diabetes medication be reduced in elderly patients? An observational study of diabetes drug withdrawal in nursing home patients with tight glycaemic control. Diabetes Res Clin Pract 2008; 82:197–202. 20. Wang ZH, Kihl-Selstam E, Eriksson JW. Ketoacidosis occurs in both Type 1 and Type 2 diabetes–a population-based study from Northern Sweden. Diabet Med 2008; 25:867–870. 21. Raijmakers NJ, van Zuylen L, Furst CJ, et al. Variation in medication use in & cancer patients at the end of life: a cross-sectional analysis. Support Care Cancer 2013; 21:1003–1011. As pointed out by Raijmakers et al. in a recent international survey on potentially inappropriate medications for dying patients, 81% of experts considered OAD to be inappropriate. 22. Vandenhaute V. Palliative care and type II diabetes: a need for new guidelines? Am J Hosp Palliat Care 2010; 27:444–445. 23. Andrews WJ, Vasquez B, Nagulesparan M, et al. Insulin therapy in obese, noninsulin-dependent diabetes induces improvements in insulin action and secretion that are maintained for two weeks after insulin withdrawal. Diabetes 1984; 33:634–642.
Volume 8 Number 4 December 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
