Utility of Postmortem Vitreous Beta-Hydroxybutyrate Testing for Distinguishing Sudden from Prolonged Deaths and for Diagnosing Ketoacidosis

A retrospective, cross-sectional analysis of vitreous beta-hydroxybutyrate (BHB) on 967 forensic cases over a two-year period was conducted. Cases were sorted into six categories of death: (i) sudden traumatic/non-natural (ST), (ii) sudden natural (SN), (iii) prolonged traumatic/non-natural (PT), (iv) prolonged natural (PN), (v) diabetic ketoacidosis (DKA), and (vi) alcoholic ketoacidosis (AKA). The mean BHB for all cases was 1.67 mmol/L (17.4 mg/dL; range: 0.11–18.02 mmol/L). The numbers of DKA, AKA, PN, PT, SN, and ST deaths were 21, 5, 155, 258, 275, and 253, respectively. Their mean vitreous BHBs were as follows: 11.04 mmol/L (DKA), 8.88 mmol/L (AKA), 1.56 mmol/L (PN), 1.55 mmol/L (PT), 1.26 mmol/L (SN), and 1.38 mmol/L (ST). There was a statistically significant difference between the mean BHBs of the PN and SN death groups (p < 0.001), as well as between those of the PT and ST death groups (p = 0.004). Given the overlapping ranges seen between the prolonged and sudden death groups, the identified differences did not hold clinical significance. In addition, we sought to determine a threshold value for vitreous BHB to definitely diagnose cases of ketoacidosis. BHB threshold concentrations between 2.5 and 5 mmol/L produced sensitivities >92% and specificities >96%. A receiver operator characteristic curve found 3.43 mmol/L to be the optimal cutoff value, demonstrating a specificity of 98.3% and a sensitivity of 96.2%.     

Armanni–Ebstein Lesions in Terminal Hyperglycemia

Armanni–Ebstein lesions (AEL) occur in deaths related to uncontrolled diabetes mellitus. To investigate the relationship between AEL and terminal hyperglycemia, we retrospectively reviewed 71 cases with vitreous glucose levels ≥11.1 mmol/L; 27 (38%) cases had AEL (vitreous glucose 14.0–77.3 mmol/L); and 44 cases (62%) did not (vitreous glucose 11.1–91.9 mmol/L). There was no significant difference (p = 0.271) in vitreous glucose levels between the cases with AEL (mean 39.2, SD 16.7 mmol/L) and those without (mean 34.2, SD 19.8 mmol/L). Similarly, there was no difference in the degree of dehydration, renal failure, or osmolality. However, there was a significantly higher level of β‐hydroxybutyrate among the cases with AEL compared to those without (p = 0.007), suggesting that ketoacidosis may facilitate the development of AEL. Given the possible synergistic role of β‐hydroxybutyrate, the correlation between AEL and terminal hyperglycemia in animal studies may not be applicable to humans. AEL may also possibly occur with sublethal elevations in glucose.     

Basal Vacuolization in Renal Tubular Epithelial Cells at Autopsy and Their Relation to Ketoacidosis

Basal vacuolization of renal tubular epithelial cells is a useful postmortem marker for ketoacidosis. To investigate its incidence and relationship to the severity of ketoacidosis, 158 autopsy cases with elevated β‐hydroxybutyrate (>1 mmol/L) over a 7‐year‐period were retrospectively reviewed. Sixty‐eight cases (43%) exhibited basal vacuolizations (vitreous β‐hydroxybutyrate: 1.16–29.35 mmol/L, mean 10.28 mmol/L), and 90 cases (57%) did not (vitreous β‐hydroxybutyrate: 1.03–13.7 mmol/L, mean 2.84 mmol/L). Quantitative analysis revealed on average a fourfold elevation in β‐hydroxybutyrate in cases with basal vacuolizations compared to those without; 10.3% of cases with β‐hydroxybutyrate concentrations between 1.01 and 2.00 mmol/L had basal vacuolizations, and this incidence increased to 33.3% with concentrations between 4.01 and 6.00 mmol/L. A marked increase in incidence to >70% was observed with concentrations >6.00 mmol/L, and basal vacuoles were invariably present (100%) with concentrations >14.01 mmol/L. This study demonstrates that basal vacuolizations are a sensitive marker for significant ketoacidosis and reaffirms its use as an indicator for likely cases of fatal ketoacidosis at autopsy.     

Postmortem vitreous Beta-hydroxybutyrate: interpretation in a forensic setting*

Abstract: Vitreous beta‐hydroxybutyrate (BHB) was retrospectively analyzed in 1795 forensic cases using the Pointe Scientific method. Comparison of vitreous BHB with vitreous glucose in 1781 of the cases showed moderately good correlation r = 0.731. Comparison with blood alcohol levels in 1561 of the cases showed no correlation r = ?0.053. Vitreous BHB was a marker of diabetic ketoacidosis when above 6.0 mM with a vitreous glucose over 200 mg/dL. It was an indicator (>50%) for alcoholic ketoacidosis when above 6.0 mM with a vitreous glucose below 200 mg/dL. Recommendations for interpretation of vitreous BHB: <0.4 mM normal; 0.41–1.2 mM slightly elevated, rarely (<1%) of concern; 1.21–2.0 mM moderately elevated, less rarely (2.5%) of concern; 2.01–6.0 mM significantly elevated, frequently of concern (12–48%); >6.0 mM usually (100% in this study) indicated life‐threatening conditions. Vitreous BHB was helpful evaluating cases with ketogenic conditions, especially diabetes and alcoholism.     

Investigation of markers to indicate and distinguish death due to alcoholic ketoacidosis, diabetic ketoacidosis and hyperosmolar hyperglycemic state using post-mortem samples

Data from 191 post-mortem cases where post-mortem blood beta-hydroxybutyrate (βHB) and acetone concentrations and vitreous humor glucose concentrations (where available) had been measured were retrospectively investigated to determine the markers required to identify and distinguish between Alcoholic Ketoacidosis (AKA), Diabetic Ketoacidosis (DKA) and Hyperosmolar Hyperglycemic State (HHS). Blood βHB concentrations above 250 μg/mL were considered significant and it was shown to be the preferred marker of ketoacidosis. All cases with significant βHB detected also had acetone present (greater than 2mg/dL) demonstrating that acetone can be used as a marker to identify ketoacidosis and can be used to indicate when βHB measurement is necessary. Vitreous humor glucose concentrations above 6.9 mmol/L were considered high and indicative of hyperglycemia prior to death. Vitreous humor glucose concentrations can be used to distinguish between DKA and ketoacidosis from other causes and to identify deaths due to HHS. The data showed that ketoacidosis can occur without a history of alcoholism or diabetes. Many diabetics are undiagnosed for many years. Therefore, DKA or HHS should be considered in sudden or unexplained deaths and glucose should be routinely measured especially in cases with risk factors for diabetes including obesity, old age, a history of mental health problems or treatment with atypical antipsychotic drugs including clozapine, olanzapine, quetiapine and risperidone.     

Postmortem detection of isopropanol in ketoacidosis

Isopropanol (IPA) detected in deaths because of diabetic ketoacidosis (DKA) or alcoholic ketoacidosis (AKA) may cause concern for IPA poisoning. This study addressed this concern in a 15-year retrospective review of 260 deaths in which concentrations of acetone and IPA, as well as their ratios, were compared in DKA (175 cases), AKA (79 cases), and IPA intoxication (six cases). The results demonstrated the frequency of detecting IPA in ketoacidosis when there was no evidence of IPA ingestion. IPA was detectable in 77% of DKA cases with quantifiable concentrations averaging 15.1 ± 13.0 mg/dL; 52% of AKA cases with quantifiable concentrations averaging 18.5 ± 22.1 mg/dL; and in cases of IPA intoxication, averaging 326 ± 260 mg/dL. There was weak correlation of IPA production with postmortem interval in DKA only (r = -0.48). Although IPA concentrations were much higher with ingestion, potentially toxic concentrations were achievable in DKA without known ingestion.     

Renal Tubular Epithelial Vacuoles—A Marker for Both Hyperlipidemia and Ketoacidosis at Autopsy

Review of 15 cases of nephrotic syndrome found that eight had significant hyperlipidemia with serum cholesterol levels ranging between 10.59 and 18.60 mmol/L (mean 12.88) and serum triglyceride levels between 2.30 and 9.92 mmol/L (mean 4.58); all of these cases displayed basal lipid vacuolization. Seven of the 15 study cases had normal–mild hyperlipidemia with serum cholesterol levels ranging between 4.71 and 7.54 mmol/L (mean 6.02) and serum triglyceride levels between 0.65 and 4.1 mmol/L (mean 1.57). Six of the seven cases had basal lipid vacuoles (86%). Of these, five cases were hyperlipidemic and one case had borderline hyperlipidemia with a serum cholesterol level of 4.71 mmol/L. Although hyperlipidemia was associated with renal tubular epithelial vacuolization, the vacuoles appeared morphologically different to those found in ketoacidosis. This study has shown that while hyperlipidemia in isolation may result in basal lipid vacuolization within renal tubular epithelial cells, the phenotype differs from that observed in ketoacidosis.     

Basal renal tubular epithelial cell vacuolization and alcoholic ketoacidosis

Subnuclear renal tubular epithelial cell vacuolization is a marker for diabetic ketoacidosis. Whether it is because of hyperglycemia or of ketoacidosis is unclear. To examine the effect of ketoacidosis on renal cells in isolation, five cases of lethal alcoholic ketoacidosis without hyperglycemia were examined (vitreous humor β-hydroxybutyrate: 6.42-8.75 mM, mean 7.66 mM; and glucose: 0.1-4.2 mM, mean 1.46 mM). Microscopic examination of the kidneys revealed basal vacuoles in three cases (60%). Seven control cases with acute alcohol toxicity without ketoacidosis (blood alcohol: 0.18-0.43%, mean 0.31%; and β-hydroxybutyrate: 0.12-0.42 mM, mean 0.21 mM) did not have these changes. In this study, basal epithelial vacuolization was found only in cases with significant ketoacidosis. Although the numbers are small, the finding of basal renal tubular epithelial vacuolization in normoglycemic cases with elevated β-hydroxybutyrate levels provide further evidence that disordered lipid metabolism may be involved in the pathogenesis of this phenomenon.     

Beta-hydroxybutyric acid--an indicator for an alcoholic ketoacidosis as cause of death in deceased alcohol abusers

We analyzed the postmortem blood of a total of 100 fatal cases for beta-hydroxybutyric acid (BHBA). In 25 cases of sudden and unexpected death of alcoholics we found pathologically increased levels of BHBA of 1260 to 47200 (median 8000) micromol/L. This led us to the diagnosis of an alcoholic ketoacidosis (AKA) as cause of death in these cases. The control group of 69 postmortem cases revealed that BHBA concentrations below 500 can be regarded as normal, and values up to 2500 micromol/L as elevated. Our study shows that BHBA values over 2500 micromol/L could lead to death, if no medical attention is sought. During storage we did not find any indication of postmortem formation or decomposition of BHBA in blood in vitro or in the corpses. In our opinion, BHBA should be considered the diagnostic marker of choice for the postmortem determination of alcoholic ketoacidosis (AKA) as the cause of death. The classical indications of such deaths are: unexpected death of a chronic alcoholic; none or only traces of ethanol in the blood; increased acetone blood concentration; and neither autopsy, histology, microbiology, nor toxicology reveal the cause of death. In six further cases a diabetic ketoacidosis (DKA) was diagnosed as the cause of death.     

How Useful is Basal Renal Tubular Epithelial Cell Vacuolization as a Marker for Significant Hyperglycemia at Autopsy?

Basal vacuolization of renal tubular epithelial cells (so-called Armanni-Ebstein phenomenon) has been attributed to hyperglycemia causing accumulation of cytoplasmic glycogen. Review of 34 autopsy cases with significant hyperglycemia (vitreous glucose ≥ 15 mmol/L/270 mg/dL) was undertaken to determine whether there was any significant association between the degree of hyperglycemia and the severity of this morphological change (graded as 0, 1+, 2+, and 3+). No association was demonstrated. Review of the subgroup of 14 cases with terminal hyperglycemia without ketoacidosis was then undertaken to assess the effect of hyperglycemia in isolation on renal tubular epithelial cells. Vitreous glucose levels in these 14 cases ranged from 17 to 49.7 mmol/L (306-894.6 mg/dL) with a mean of 26.25 mmol/L (472.5 mg/dL) and β-hydroxybutyrate levels ranged from 0.02 to 2.55 mmol/L (0.36-45.9 mg/dL) with a mean 0.79 mmol/L (14.22 mg/dL). Not one of the latter cases displayed basal vacuolization. No relationship between basal vacuolization of renal tubular epithelial cells at autopsy and terminal hyperglycemia could, therefore, be demonstrated.     

Fatal Diabetic Ketoacidosis—A Potential Complication of MDMA (Ecstasy) Use

A 19‐year‐old woman with insulin‐dependent diabetes mellitus was found dead in bed having allegedly recently taken ecstasy and consumed alcohol. At autopsy, there were microhemorrhages in the brain with subnuclear vacuolization and Armanni–Ebstein changes in renal tubules. Biochemical analyses confirmed diabetic ketoacidosis (vitreous glucose—46.5 mmol/L; β‐OH butyrate—13.86 mmol/L.). Toxicological analyses of blood showed a low level of 3,4‐methylenedioxy‐methamphetamine (MDMA) (0.01 mg/L), with acetone but no alcohol or other common drugs. Death was attributed to diabetic ketoacidosis most likely provoked by mixed MDMA/alcohol ingestion. Although the use of illicit drugs by young individuals with diabetes mellitus is being increasingly recognized, it has been noted that there is minimal information about the relationship between drug use and acute diabetic complications. Toxicological screening of cases of lethal diabetic ketoacidosis in the young may clarify lethal mechanisms in individual cases and also help to determine the extent of this problem.     

Renal Cortical Pallor—A Useful Macroscopic Marker for Metabolic Derangements at Autopsy

Renal cortical pallor was studied as a potential marker at autopsy of diabetic ketoacidosis in 23 cases, hyperglycemic nonketotic coma in eight cases, and alcoholic ketoacidosis in five cases (vitreous humor glucose level ≥11.1 mM; β‐hydroxybutyrate level ≥5 mM). Renal cortical pallor was noted on macroscopic examination in 10 of 23 cases of lethal diabetic ketoacidosis (43.5%), three of eight cases of fatal hyperglycemic nonketotic coma (37.5%), and in two of five cases of alcoholic ketoacidosis (40%). Histologic examination revealed basal vacuolization of renal tubular epithelial cells in 12 cases, Armanni–Ebstein lesions in 10, and osmotic nephrosis in three. Although renal cortical pallor did not appear to be a particularly sensitive marker for hyperglycemia or ketoacidosis, and did not correlate with the severity of these parameters, it may still represent a useful macroscopic marker for underlying metabolic conditions at autopsy and should therefore prompt measurement of vitreous humor glucose and β‐hydroxybutyrate levels.     

Septic Ketoacidosis—A Potentially Lethal Entity with Renal Tubular Epithelial Vacuolization

Fatal ketoacidosis due to diabetes mellitus, alcoholism, and starvation may produce characteristic basal vacuolization of renal tubular epithelial cells (RTEC). Septic ketoacidosis has recently been recognized clinically as a distinct condition in which septicemia can lead to elevation of ketones and various anions unrelated to diabetes mellitus, alcoholism, or caloric deprivation. We report four lethal cases with significantly elevated vitreous ketones secondary to sepsis and/or severe localized infection in individuals with no history of diabetes mellitus, alcoholism, or starvation. Three of four cases exhibited typical basal vacuolization of RTEC. We suggest that septic ketoacidosis is an appropriate cause of death in the forensic setting where sepsis or severe localized infection is found with significant ketoacidosis (β‐hydroxybutyrate > 5 mmol/L)—in the absence of diabetes mellitus, alcoholism, starvation, or other states associated with accelerated ketogenesis. The finding of basal vacuolization of RTEC in such cases provides morphological support for the underlying metabolic derangement.     

Urinary endogenous concentrations of GHB and its isomers in healthy humans and diabetics

Urinary endogenous concentrations of gamma-hydroxybutyric acid (GHB), alpha-hydroxybutyric acid (AHB) and beta-hydroxybutyric acid (BHB) have been investigated for both healthy humans and diabetics by using a newly optimized GC-MS procedure. The endogenous concentrations in healthy volunteers' urine ranged 0.16-2.14 microg/ml for GHB, 0.10-2.68 microg/ml for AHB and 8.51-34.7 microg/ml for BHB. In diabetics, the concentrations ranged 0.17-3.03 microg/ml for GHB, 0.14-124 microg/ml for AHB and 4.94-4520 microg/ml for BHB. Although notably elevated BHB and AHB concentrations were observed for severely uncontrolled diabetics, their GHB concentrations ranged within or near the range seen in healthy humans. The results of this study confirm the previously suggested 10 microg/ml cutoff concentration of urinary GHB to distinguish exogenous GHB, even for uncontrolled diabetic patients suffering severe ketoacidosis.     

The Etiology of Basal Vacuolizations in Renal Tubular Epithelial Cells Evaluated in an Isolated Perfused Kidney Model

To determine whether basal lipid vacuolization characteristic of ketoacidosis could be induced with short‐term hypertriglyceridemia, adult Sprague Dawley rat kidneys were perfused in an isolated perfused kidney model with, and without, 11.3 mM (10 g/L) of triglycerides in Krebs‐Henseleit buffer, for 1 and 2 h (n = 5/group). Additional treatments included perfusion with triglycerides with 20 mM of β‐hydroxybutyrate and 2 mM of acetoacetate (n = 5) and perfusion with triglycerides with 70 mM of glucose (n = 1). Basal vacuolization was produced in all groups, but differed in morphology to that reported in postmortem studies. There was no further increase in vacuolization after 2 h of perfusion compared to 1 h (p = 0.24), and the addition of ketones did not alter the morphology or extent of vacuolization. This study using an ex vivo model has confirmed that isolated hypertriglyceridemia is sufficient to cause basal lipid vacuolization in renal tubular epithelial cells, but with different morphology to vacuoles observed in lethal ketoacidosis at autopsy.     

Fatality from olanzapine induced hyperglycemia

A case history of a 31-year-old male schizophrenic patient is presented. The man was treated with olanzapine for three weeks before he died. After one week on a 10 mg daily dose of olanzapine, his fasting blood glucose was elevated to 11.3 mmol/L (203 mg/dL). In order to treat more aggressively his psychosis, the olanzapine dose was raised to 20 mg daily resulting in his fasting blood glucose climbing to 15.8 mmol/l (284 mg/dL). On the days preceding his death, he became progressively weaker, and developed polydipsia with polyuria. He had no personal or family history of diabetes mellitus and he was on no other medication at the time of his death. Postmortem blood, vitreous humor, and urine glucose concentrations were 53 mmol/L (954 mg/dL), 49 mmol/L (882 mg/dL), and 329 mmol/L (5922 mg/dL), respectively. Drug screen on urine and blood indicated only a small amount or olanzapine and no alcohols. Peripheral blood olanzapine concentration was within therapeutic limits, 45 ng/mL. Analysis of vitreous humor and urine revealed severe dehydration with small amounts of ketones. Death was attributed to hyperosmolar nonketotic diabetic coma, and olanzapine was felt most likely to be the cause. Another atypical neuroleptic, clozapine, has also been associated with the development and exacerbation of diabetes mellitus or diabetic ketoacidosis. We recommend including vitreous glucose and beta-hydroxybutyrate analysis as part of postmortem toxicology work up when the drug screen reveals the presence of either olanzapine or clozapine.     

Vitreous humor magnesium in alcoholics

Postmortem vitreous humor magnesium concentrations were determined in 27 alcoholics and in 27 nonalcoholics who died of various causes. No significant differences in magnesium values were present between these two groups. The concentrations in those dying with alcohol withdrawal syndrome or in those with postmortem findings limited to hepatic fatty change did not differ significantly from those in the nonalcoholic group or from those of the alcoholics dying of other causes. There was no correlation between vitreous humor magnesium concentrations and postmortem interval in any group. We conclude that vitreous humor magnesium determination is not helpful in the postmortem evaluation of suspected alcohol-related deaths or in determining postmortem interval.     

Measurement of chemical analytes in vitreous humor: stability and precision studies

The analytic accuracy and precision for measurement of chemical analytes in vitreous humor (VH) are critical if results are to be used in forensic pathology. The purpose of our study was to demonstrate the stability and the reproducibility of VH sodium, potassium, chloride, glucose, urea nitrogen, acetone, and beta-hydroxybutyrate in specimens obtained from both eyes in medical examiner cases. We also compared with calculated VH osmolalities. Small but significant increases were observed in VH electrolyte concentrations in specimens refrigerated 6-12 months: sodium pre 144 mmol/L, post 151 mmol/L; potassium pre 12.0 mmol/L, post 12.8 mmol/L; chloride pre 121 mmol/L, post 123 mmol/L. No differences were observed between eyes, and within-day precision for all electrolyte measurements were excellent, (<1%). Frozen specimens showed significantly higher measured (439 mOsmol/kg) as compared with calculated osmolality (305 mOsmol/kg), with 1% within-day precision and no significance between eye variation for glucose and urea nitrogen. In 20 of 24 medical cases selected for possible ketoacidosis, measurement of beta-hydroxybutyrate concentrations appears to be a promising diagnostic biomarker for confirming suspected ketosis in medical examiner cases by means of VH analysis.     

Postmortem Distribution of 3‐Beta‐Hydroxybutyrate

The concentrations of 3‐beta‐hydroxybutyrate (3HB) in femoral blood, urine, vitreous humor as well as pericardial and cerebrospinal fluids were retrospectively examined in a series of medico‐legal autopsies, which included cases of diabetic ketoacidosis, hypothermia fatalities without ethanol in blood, bodies presenting mild decompositional changes, and sudden deaths in chronic alcoholics. Similar increases in 3HB concentrations were observed in blood, vitreous, and pericardial fluid, irrespective of the cause of death, suggesting that pericardial fluid and vitreous can both be used as alternatives to blood for postmortem 3HB determination. Urine 3HB levels were higher than blood values in most cases. Cerebrospinal fluid 3HB levels were generally lower than concentrations in blood and proved to be diagnostic of underlying metabolic disturbances only when significant increases occurred.