The patient has a history of bronchiectasis involving left and right lower lung, restrictive lung disease due to old pulmonary tuberculosis with component of diabetic or uremic neuropathy on nocturnal non-invasive ventilation since October 2011, diabetes mellitus on insulin, hypothyroidism on thyroxine replacement, diabetic nephropathy and end stage renal failure on helper continuous ambulatory peritoneal dialysis (CAPD) since February 2007. The helper CAPD was performed by his wife who also has a history of end stage renal disease and she had received renal transplantation in mainland China.
Patient was admitted on 31st January 2012 through the Accident and Emergency Department of a public hospital for persistent abdominal pain on admission day. The epigastric pain was not related to food intake. There was no vomiting or diarrhea. Bowel opening on admission day was 4 times and it was all formed stool. On admission he has no fever and temperature was 37.2 Degree Celsius. Blood pressure was 123/94 mmHg and pulse rate was 110 per minute. Oxygen saturation was 98% in room air and respiratory rate was 16 per minute. Physical examination showed soft abdomen with tenderness over the upper abdomen, no guarding, no obvious rebound tenderness, and active bowel sound. Tenckhoff catheter exit site was well. Tenckhoff catheter was surgically inserted at the abdominal wall to facilitate the inflow and outflow of dialysate fluid. Chest was clear on examination and the cardiovascular examination showed normal heart sound with no murmur. His initial white cell count is mildly elevated to 11.2 x 10^9/L, hemoglobin 8.6 g/dL (low), and platelet 548 x 10^9/L (elevated). For the renal function test, urea is elevated to 23.7 mmol/L, creatinine 607 umol/L (markedly elevated) and in the range of end stage renal disease, sodium 128 mmol/L (low), and potassium 3.1 mmol/L (low). For the liver function test, the alkaline phosphatase is 132 IU/L (elevated), and albumin 22 g/L (low). The elevated liver enzyme and low albumin level indicate liver injury and poor liver synthetic function. The calcium level is 2.03 mmol/L (low) but the corrected level is 2.48 mmol/L and it is within the normal limit. The phosphate level is 1.22 mmol/L(normal). Amylase is less than 5 IU/L (normal). CXR showed no free gas under diaphragm with the bilateral pleural effusion same as before when compared with old Chest X-ray film. The CAPD dialysate was found to be turbid. Peritoneal dialysate fluid red cell count <10/cmm, total nucleated cell is 1340/cmm, neutrophil is 93%, lymphocyte is 1%, and monocyte is 6%. The peritoneal dialysate fluid report commented the presence of cell clot and cell clump which might indicate unreliable count. Peritoneal dialysate for culture grew Pseudomonas aeruginosa which is sensitive to gentamicin, and cefoperazone plus sulbactam. Patient was diagnosed to have CAPD peritonitis. He was put on intra-peritoneal antibiotics including gentamicin and cefazolin on 31st January 2012. After initiation of antibiotics, patient had remained afebrile, alert and conscious. Glasgow coma scale was full and daily activities remained independent. He could walk unaided in ward. Vital signs monitoring in general ward show oxygen saturation remained around 94-99% in room air. Blood pressure had remained stable after admission with systolic blood pressure ranged between 90 and 123 mmHg. Diastolic blood pressure remained around 50 to 70 mmHg. Pulse rate showed downward trend from 110 per minute to 90 per minute. His blood sugar level was stable during hospitalization. Insulin dose was not adjusted after admission. Abdominal pain subsided on 1st February 2012 and he had no shortness of breath. Repeated peritoneal dialysate fluid cell count on 1st February 2012 showed red cell count <10/cmm, total nucleated cell is 450/cmm (downward trend), neutrophil is 86% (improving) and monocyte is 14%. In view of patient’s improved condition, my senior had decided to allow the patient home and continue home intra-peritoneal antibiotics by community nursing service (CNS) in the morning during senior round. Patient was keen for discharge after the discussion, and he had agreed to the option of home intra-peritoneal antibiotics by CNS due to poor eye-sight. Renal nurse had later assessed the patient in the afternoon, and she had acknowledged me about the turbidity of the dialysate fluid. However in view of the prior agreed discharge plan during senior round with the patient, patient was discharged on 1st February 2012. He could leave the hospital independently.
On 2nd February 2012, the patient was admitted to the Accident and Emergency Department of the public hospital with collapse at 8:41 am. He was found to have sweating and dullness in the morning by wife at 8:05 am. Wife had given him some sugar drinks then she went out to buy breakfast. When she returned home, patient was found to be unconscious and she called the ambulance afterwards at 8:30 am. Cardiopulmonary resuscitation performed by ambulance staff. He was found to have hemostix (instant blood glucose check by finger pricking) of 2.0 at that time. There is no record of hypoglycemia correction on ambulance. On arrival to the Accident and Emergency Department, patient was resuscitated for cardiac arrest. Patient was intubated, adrenaline was given, and intravenous 50% dextrose of 40ml was given at 8:51 am as the hemostix was 2.8 on arrival to the Emergency. The hemostix was corrected to 12.2. However, due to hyperkalemia (K: 6.67), he received 10 units of intravenous insulin actrapid and 20 ml of 50% dextrose at 8:55 am. Chest X-ray showed white out of left lung which was similar to previous old Chest X-ray films. The intensive care unit (ICU) was consulted at the Emergency for post cardiac arrest resuscitation. However patient was not admitted to ICU. Patient was transferred to medical general ward for further management. He was put on double inotropes with dopamine and adrenaline infusion. He was given intravenous antibiotics meropenem and he was supported with intravenous dextrose infusion due to recurrent and persistent hypoglycemia. His hemostix was found to be <1.1 at 14:20 and 40 ml of 50% dextrose was given at 14:25. 10% dextrose infusion was started at 14:30. Dextrose solution was changed to 20% when the hemostix was found to be 2.4 at 16:35.
His subsequent investigation results showed normal white cell count of 7.4 x 10^9/L, Hemoglobin 8.6 g/dL (low and static) and platelet 518 x 10^9/L (elevated). The renal function test showed urea of 20.3 mmol/L, creatinine 639 umol/L, sodium 133 mmol/L, and potassium 3.1 mmol/L. The liver function test showed alkaline phosphatase of 169 IU/L (increased), albumin 19 g/L, and aspartate transaminase 159 IU/L (grossly elevated). The calcium is 2.5 mmol/L (normal) but the corrected calcium was 3.025 mmol/L (elevated) and phosphate is 3.6 mml/L (elevated). The elevation of corrected calcium is probably the result of 10 ml of calcium chloride injection at 8:53 am for cardiac protection because of initial instant result of hyperkalemia. For the cardiac enzymes, the lactate dehydrogenase is 431 IU/L (elevated), creatinine kinase is 458 IU/L(elevated), and Troponin T-high sensitivity is 288 ng/L (elevated). Arterial blood gas showed pH 7 (low), partial pressure of carbon dioxide is 95 mmHg (elevated), partial pressure of oxygen is 144 mmHg (elevated), bicarbonate is 23 mmol/L (normal), and base excess is -10 mmol/L. Electrocardiogram showed sinus rhythm and heart rate was 137 per minute on 31st January 2012. A new onset T wave inversion was found in leads I, II, aVF, V5-6 on 2nd February 2012. The clinical picture suggests post cardiac arrest myocardial ischemia and tissue ischemic injury. His condition remained poor with systolic blood pressure around 90 to 100 mmHg and diastolic blood pressure around 40 to 50 mmHg, oxygen saturation 96% on 100% oxygen supplement, and afebrile. Patient developed cardiac arrest on the same day of admission on 2nd February 2012 in the evening, ‘Do Not Resuscitate’ was agreed by wife after discussion with on call medical officer and patient was certified death at 18:42 pm on 2nd February 2012.
Hypoglycemia as a result of excessive insulin injection
The patient has history of multiple admissions due to hypoglycemia. He was admitted on 5th February 2011 for hypoglycemia due to disrupted CAPD schedule. His usual CAPD regimen at that time was one bag of 7.5% dextrose dialysate and three bags of 2.5% dextrose dialysate per day. He had increased work load before Chinese New Year, skipped one bag of 2.5% dextrose dialysate and extended the indwelling time of the 7.5% dextrose dialysate. Despite this wrong regimen he still keeps the usual dose of insulin. He contacted renal hotline on 28th February 2011 for hypoglycemia and was encouraged to take snacks before going to bed. He was admitted on 8th June 2011 for hypoglycemia precipitated by sepsis. On 18th July 2011, he was admitted for hypoglycemia after adjustment of insulin. All these episodes of hypoglycemia are correctable and patient was discharged subsequently.
During the admission on 2nd February 2012, his persistent and recurrent hypoglycemia was not due to sepsis. His clinical condition is improving as evidenced by the investigation parameters. His initial white cell count was 11.2 (mildly elevated) on 31st January 2012 but was 7.4 (normal) on 2nd February 2012. The peritoneal dialysate analysis on 31st January 2012 showed fluid red cell count <10, total nucleated cell 1340, neutrophil 93%, lymphocyte 1%, and monocyte 6%. Repeated peritoneal dialysate analysis on 1st February 2012 showed red cell count <10, total nucleated cell 450 (downward trend), neutrophil 86% (downward trend), and monocyte 14%. There was no fever. Blood pressure had remained stable after admission on 31st January 2012 and 1st February 2012 with systolic blood pressure ranged between 90 and 123 mmHg. Diastolic blood pressure remained around 50 to 70 mmHg. Pulse rate showed downward trend 110 per minute to 90 per minute. Blood pressure was similar to previous hospital admission on 18th January 2012 which was 80/60 mmHg and pulse rate of 70 per minute. CAPD regimen was not adjusted upon discharge. Patient can leave the hospital independently. Community nursing service was arranged for him since the patient cannot inject intra-peritoneal antibiotics for himself due to poor eye sight. The persistent hypoglycemia on 2nd February 2012 precipitates the cardiac arrest due to underlying poor heart function. Echocardiogram on 19th January 2012 showed poor left and right ventricular function and ejection fraction of 20%.
Patient’s hypoglycemia was a result of insulin isophane injection because actrapid is short-acting and it is not possible for its effect to last from 8:55 am till the hypoglycemia happened at 14:20 and 16:35 when giving intravenously. According to British National Formulary (1), actrapid is soluble insulin with rapid onset of action (30 to 60 minutes), a peak action between 2 and 4 hours, and duration of action up to 8 hours when giving subcutaneously. When injected intravenously, soluble insulin has a very short half-life of only about 5 minutes and its effect disappears within 30 minutes. According to British National Formulary (2), insulin isophane is a suspension of insulin with protamine and it is intermediate-acting and from its duration of action it can explain the persistent hypoglycemia in the afternoon when the injection was made in the morning subcutaneously. When given by subcutaneous injection, intermediate-acting insulin has an onset of action of approximately 1-2 hours, a maximal effect at 4-12 hours, and a duration of 16-35 hours.
The usual dose of insulin for the patient is protaphane 12 units in the morning subcutaneously. During admission on 31st January 2012, his blood glucose control was stable on insulin and there was no hypoglycemia with regular meals at the hospital. On 31st January 2012, insulin isophane was given at 8:00 am. Hemostix levels were 18.9 at 7:40 am, 13.6 at 11:30 am, 12.6 at 16:45, and 17.6 at 20:38. On 1st February 2012, insulin isophane was given at 7:30 am and hemostix levels were 4.3 at 6:43 am and 10.3 at 12:00 noon. There was no adjustment of insulin dose during the same admission. With usual dose of insulin, the hypoglycemia should be correctable with sugar drinks or intravenous glucose and the hypoglycemia will not become persistent and recurrent in the afternoon. The only explanation is excessive dose of insulin injected subcutaneously in the morning at home. We need to consider the possibility of insulin injection by others since the patient has poor eye sight.
The time of making injection is an important consideration in this suspected insulin murder case. In this case the patient was injected with massive dose of subcutaneous insulin and the action profile in massive insulin overdose will be different from that of the usual dose. A case report (3) stated that a patient who had taken a large dose of neutral protamine Hagedorn (NPH) insulin and the highest level of circulating insulin was measured 6 hours after administration of the NPH insulin. Another study (4) by the same authors stated that a 20-year-old diabetic patient was injected with over-dose of isophane insulin suspension shortly before hospitalization and insulin was peaked at 6 hours and the hypoglycemic symptoms happened at 6 hours as well. For this patient, his lowest hemostix level is <1.1 at 1420 and counting backward for 6 hours, the injection time will be at 0820 on 2nd February 2012 that is after the presence of hypoglycemic symptoms. It was mentioned in the admission notes from the Accident and Emergency Department that the patient has decreased appetite and this might also contribute to the presence of hypoglycemia in the morning. Also, there is one episode of borderline hemostix level of 4.3 at 6:43 am on 1st February 2012 and he has a tendency to be hypoglycemic even on usual dose of insulin. If this is true, it is improbable for the patient to inject insulin for himself in the presence of hypoglycemic symptoms provided that the peak insulin level of massive insulin overdose happened at 6 hours and the injection must be made by others and it will very likely to be murder then. In a retrospective analysis of 243,222 patients (5), the incidence of hypoglycemia was significantly higher among patients with chronic kidney disease (defined as estimated glomerular filtration rate <60mL/min per 1.73 m2) compared with patients without chronic kidney disease both among those with diabetes (10.72 versus 5.33 per 100 patient-months, respectively) and without diabetes (3.46 versus 2.23 per 100 patient-months, respectively).
Insulin has a molecular weight of 6000. Of the total renal insulin clearance, approximately 60 percent occurs by glomerular filtration and 40 percent by extraction from the peritubular vessels. Insulin in the tubular lumen enters proximal tubular cells by carrier-mediated endocytosis and is then transported into lysosomes where it is metabolized to amino acids (6). The renal clearance of insulin is 200 mL/min, significantly exceeding the normal glomerular filtration rate (GFR) of 120 mL/min due to the contribution of tubular secretion. From this rate of renal clearance, it can be calculated that six to eight units of insulin are degraded by the kidney each day, which accounts for approximately 25 percent of the daily production of insulin by the pancreas. The contribution of renal metabolism is enhanced in diabetic subjects receiving exogenous insulin, since injected insulin enters the systemic circulation directly without first passing through the liver (7). The dose of insulin should be reduced by as much as 50 percent when the GFR is less than 10 mL/min (8). Because of a large molecular weight (6000 daltons), insulin is not significantly removed by peritoneal dialysis or hemodialysis (9). From the above evidence, we can calculate that the insulin production in normal subjects is 24 to 32 units per day. If the exogenous insulin required for a diabetic patient with normal renal function is similar, the dose should be reduced to 12 to 16 units in a patient with end stage renal disease with GFR less than 10 mL/min as our patient. The patient was on regular dose of 12 units of insulin which is the correct insulin requirement for him. As a result, the cause of persistent and recurrent hypoglycemia in his case must be due to excessive insulin injection. Based on a study in Japan (10), if a huge amount of insulin is subcutaneously injected, the duration of hypoglycemia depends on the dose of insulin rather than the type of insulin. There was a significant correlation between recovery time from hypoglycemia and insulin dose and this correlation was expressed as y=0.045x; where y is time (h) and x is insulin dose (U). The patient did not recover from hypoglycemia when the subcutaneous insulin was injected at 0820 am. He passed away at 1842 pm on the same day so that y is longer than 10 hours. From the equation, the insulin dose will be more than 222 units.
The patient has end stage renal disease and the question is whether the renal disease will affect the recovery time from insulin overdose. A case report (11) in Japan reported a 57-year-old Type 2 diabetic man with chronic renal failure under hemodialysis and he was injected with 1,200 units of regular insulin subcutaneously. He was admitted to the hospital two hours after the overdose injection because of severe hypoglycemia and unconsciousness. Despite continuous glucose injection, hypoglycemia persisted for twenty-four hours and he recovered completely without any complications. From the equation above the estimated recovery time will be 54 hours with 1,200 units of insulin injection and the insulin dose will be 533 units if the recovery time is twenty-four hours. It showed that this equation is more conservative than we think and the patient could be injected with even higher dose. Normal deduction will predict that the recovery time will be prolonged in end stage renal disease but this case report showed that the recovery time was not affected by the renal function. Usual end stage renal disease patients need insulin dose adjustment only as illustrated above. This does not apply to the patient because patient was injected with excessive dose.
The possibility of insulin injection via intra-peritoneal route
We can rule out the possibility of intra-peritoneal insulin injection based on the following reasons:-
The patient was not usually on intra-peritoneal insulin so that the helper did not know how to inject insulin into the dialysis solution container.
The patient needs pre-arranged community nursing service for intra-peritoneal antibiotic injection daily during the period of CAPD peritonitis so that he does not know how to perform intra-peritoneal injection.
CAPD peritonitis does not alter the insulin requirements (12) but the problem is relatively long, 3.8 cm (1.5 in) needles should be used to ensure that the full dose of insulin is injected into the dialysis solution container rather than being trapped in the infusion port (13). Without this instrument the insulin injection will not achieve its effect.
Mechanism of iatrogenic hypoglycemia induced sudden cardiac death
According to an article written by Professor PE Cryer (14), the Irene E. and Michael M. Karl Professor of Endocrinology and Metabolism in Medicine at the Washington University School of Medicine, iatrogenic hypoglycemia causes hypoglycemia-associated autonomic failure (HAAF). The components of HAAF are defective glucose counter-regulation (a result of absent decrements in insulin, absent increments in glucagon, and attenuated increments in epinephrine), hypoglycemia unawareness (a result of attenuated increments in sympathoadrenal activity), and decreased baroreflex sensitivity. Cardiovascular HAAF that is decreased baroreflex sensitivity results in increased vulnerability to ventricular arrhythmia and subsequently sudden death. Defective glucose counter-regulation and hypoglycemia unawareness lead to recurrent hypoglycemia and sympathoadrenal activation which in turn cause prolonged corrected QT interval of the heart, increase in intracellular calcium, decrease in potassium, increase in oxygen consumption, increase in platelet function, increase in coagulation, increase in cytokines and decrease in vasodilation. With these effects, it causes ventricular arrhythmia and sudden death. This is the biochemical mechanism of action of hypoglycemia in causing cardiac arrest and it also explains why the patient develops cardiac arrest after injection with excessive dose of insulin.
Cases of insulin murder
The first (15) documented case of murder by insulin happened when Kenneth Barlow of the Bradford, Yorkshire of the Northern England discovered his wife, Elizabeth Barlow, unconscious in the bath on 4 May 1957. The Police searched his home and found a couple of used syringes in the kitchen but did not find vials of insulin or other injectable medications. Two hypodermic injection sites were identified on each buttock on 8 May 1957. Measurable quantities of insulin were found in tissues taken from Elizabeth’s buttock. Kenneth was charged with murder but he denied but he could not explain the insulin found inside Elizabeth’s body apart from suggesting that she injected herself. The improbability of self-injection into the buttock brings Kenneth Barlow to a sentence of life imprisonment on 13 December 1957. He was released from the prison 26 years later.
The second case (16) related to the insulin murder I would like to bring your attention to happened in the United States of America. The murderer named William Dale Archerd (also known as James Lynn Arden). He was found guilty on 15 March 1968 of murdering his nephew and two of his seven wives. He presumably stole or bought the insulin as the weapon to kill. He used to work as salesman, mental hospital attendant and serial matrimonialist. For much of the time he was unemployed. It begins when he contacted the police on 24 July 1956 to report an alleged robbery at his home in Los Angeles. He said the robber was armed with guns and hypodermic syringes and used it to inject him and his third wife, Zella, in the buttocks. Zella later turned into coma and convulsion and died next day without regaining consciousness. The police searched the house and found a hypodermic needle and a half-used vial of long-acting insulin in the bathroom. Zella’s death was attributed to bronchopneumonia because no poisonous substances found in her body and there was no method of measuring insulin available. The suspicion of murder remained. The second death was William’s fifth wife, Juanita. She was found to be in coma on 12 March 1958. She was taken to hospital but died only a few hours after being found and without regaining consciousness. She had a low blood sugar without obvious medical explanation. Her death was almost certainly due to insulin but the crucial test was not done. The death of his sixth wife’s ex-husband on 17 March 1960 drew the Police attention once again. On 21 August 1961, William had taken his nephew to the hospital in a semi-comatose state. His nephew was involved in hit-and-run traffic accident in Nevada. His nephew turned into fully comatose state during the evening and night of 23 August, shortly after a visit from William and he never regained consciousness and died 10 days later on 2 September 1961. The only laboratory abnormality recorded in his clinical notes was a low glucose concentration in his cerebrospinal fluid. In view of Kenneth Barlow’s case in the United Kingdom, the investigator had considered the coma was due to insulin. But they had no means of proving it and the death was put down to the car accident that had brought him into hospital. William‘s seventh wife, Mary Brinker Arden, became bankrupt one year after their marriage. They separated and when Mary had a traffic accident on 28 October 1966, William went to her house to console her and two days later Mary was found coma and admitted to the hospital and she died next day without recovering consciousness. Blood tests showed she had a very low blood glucose level. With the development of the new technology, immunohistology, they were able to identify abnormally high quantities of insulin from Mary’s brain tissue. William was found guilty by the presiding judge on three counts of first-degree murder for the deaths of his nephew, Zella and Mary on 15 March 1968 at the age of 55. He was sentenced to death. In 1972, William’s sentence was commuted to life imprisonment after US Supreme Court had ruled that the death penalty was a ‘cruel and unusual punishment’. He died of pneumonia at the age 65.
The possibility of hyperkalemia induced cardiac arrest
The instant checking of potassium at the Accident and Emergency Department on 2nd February 2012 was 6.67 mmol/L at 8:50 am and he was treated with dextrose-insulin drip afterwards. Also, it is relatively contraindicated to give insulin in a patient presented with hypoglycaemia. It is improbable that the cause of cardiac arrest is due to hyperkalemia based on following reasons:-
The result of the instant checking machine (Roche Cobas b123 POC System) is unreliable as on arrival to the ward the patient’s potassium level was 3.1 mmol/L at 10:31 am on 2nd February 2012 which was same as the potassium level on 31st January 2012. The use of instant checking machine requires frequent calibration and if the sample is not properly taken, it will lead to misleading results. The use of short acting insulin actrapid at 8:55 am will only temporarily shift the potassium into the cellular space. Since the action of actrapid given via intravenous route can only last for 30 minutes, if the potassium is really elevated on arrival to the Accident and Emergency Department, the laboratory potassium level on arrival to ward should remain high at 10:31 am instead of the actual finding of low potassium level of 3.1 mmol/L. No other potassium lowering agent was given at the Emergency Department besides insulin actrapid.
There is no way for the non-medical personnel to obtain intravenous potassium and it is improbable for an untrained person to inject the potassium via intravenous or intra-peritoneal route.
There were no electrocardiographic (ECG) changes of hyperkalemia in this case. ECG changes in hyperkalemia include tall peaked T waves with a shortened QT interval. As the hyperkalemia gets more severe, there is progressive lengthening of the PR interval and QRS duration, the P wave may disappear, and ultimately the QRS widens further to a sine wave pattern. Ventricular standstill with a flat line on the ECG ensues with complete absence of electrical activity (17). ECG of the patient showed atrial tachycardia and new onset T wave inversion in leads I, II, aVF, V5-6 at 8:58 am on 2nd February 2012.
From patient’s drug list, there is drug interaction between hydroxyzine hydrochloride and potassium chloride sustained release tablet. Hydroxyzine is the anti-histamine used to relieve patient’s skin itchiness and its anti-cholinergic effect will slow the bowel movement and increase absorption of the potassium chloride. The patient was put on potassium chloride since 19 January 2012 due to persistent hypokalemia. However, on his admission on 31st January 2012 his potassium was 3.1 mmol/L though he had taken this drug combination for 13 days. It is improbable for the potassium to rise suddenly over one day because of this drug combination.
From the above interpretation, the elevated potassium level checked instantly at the Accident and Emergency Department is probably unreliable and is not a cause of his cardiac arrest.
This is a suspected murder case and patient died from sudden cardiac death as a result of excessive dose of subcutaneous insulin leading to hypoglycemia. There are people killed by insulin as illustrated above.
British National Formulary 57 March 2009: 370.
British National Formulary 57 March 2009: 372.
Bauman WA, Yalow RS. Insulin as a lethal weapon. J Forensic Sci 1981; 26:594-98.
Bauman WA, Yalow RS. Hyperinsulinemic hypoglycemia. Differential diagnosis by
determination of the species of circulating insulin. JAMA 1984;252:2730-2734.
Moen MF, Zhan M, Hsu VD, et al. Frequency of hypoglycemia and its significance
in chronic kidney disease. Clin J Am Soc Nephrol 2009; 4:1121.
Carone FA, Peterson DR. Hydrolysis and transport of small peptides by the
proximal tubule. Am J Physiol 1980; 238: F151.
Palmer BF, Henrich WL. Carbohydrate and insulin metabolism in chronic kidney
disease. 2012 UpToDate.
Aronoff GR, Berns JS, Brier ME, et al. Drug prescribing in renal failure: Dosing
guidelines for adults, 4th ed. Philadelphia: American College of Physicians, 1999.
Insulin regular: Drug information. 2012 UpToDate.
Ohyama T, Saisho Y, Muraki A, Kawai T, Itoh H. Prediction of recovery time from
hypoglycemia in patients with insulin overdose. Endocr J 2011; 58(7):607-11.
Epub 2011 Apr 26.
Kasumi Sato, Mitsuko Iwazaki, Yukie Ito, Mikio Wakasa. Prolonged hypoglycemia
after regular insulin overdose in a diabetic patient with chronic hepatitis C undergoing hemodialysis. Nihon Toseki Igakkai Zasshi Vol. 45 (2012) No. 2 P 175-178.
Mactier RA, Moore H, Khanna R, Shah J. Effect of peritonitis on insulin and glucose absorption during peritoneal dialysis in diabetic rats. Nephron. 1990; 54(3):240-4.
Berns JS. Management of hyperglycemia in diabetics with end-stage renal disease. 2012 UpToDate.
Cryer PE. Death During Intensive Glycemic Therapy of Diabetes: Mechanisms and Implications. Am J Med 124:993-996, 2011.
Marks V, Richmond C. Kenneth Barlow: the first documented case of murder by insulin. J R Soc Med 2008: 101: 19-21.
Marks V, Richmond C. William Archerd: a serial matrimonialist killer. J R Soc Med 2008: 101: 63-66.
Mount DB. Clinical manifestations of hyperkalemia. 2012 UpToDate.