ORIGINAL ARTICLE
Clinical Course of Diabetic Ketoacidosis in Hypertriglyceridemic Pancreatitis Dania Lizet Quintanilla-Flores, MD,* Erick Joel Rendón-Ramírez, MD,* Perla Rocío Colunga-Pedraza, MD,* Jesús Gallardo-Escamilla, MD,† Sergio Antonio Corral-Benavides, MD,† José Gerardo González-González, PhD,‡§ and Héctor Eloy Tamez-Pérez, PhD§ Objectives: Hypertriglyceridemic pancreatitis (HP) is an uncommon condition accounting for 1% to 4% of cases of acute pancreatitis, mostly associated with poor glycemic control. Diabetic ketoacidosis (DKA) may complicate the clinical course of HP. Our objective was to identify clinical and demographic differences between HP and DKA patients compared with those without DKA. Methods: Fifty-five patients with HP were included. Diabetic ketoacidosis was diagnosed in 8 patients. We analyzed the severity, hospital stay, delay in oral intake, duration of insulin infusion, complete blood cell count, and triglyceride levels. Results: Diabetic ketoacidosis was associated with a more severe HP. There were no differences in hospital stay, delay in oral intake, or duration of insulin treatment in both groups. Serum amylase, lipase, and triglyceride levels were similar. Previous diagnosis of diabetes mellitus, higher Ranson and APACHE II scores, and higher serum glucose level at admission were the only predictive risk factors for DKA and HP. Conclusions: Coexistence of DKA does not modify the clinical course of HP, although a more severe episode of HP in DKA patients. Diabetic ketoacidosis was associated with higher insulin doses, without impact in triglyceride levels. Diabetic ketoacidosis and HP should be considered when a previous diagnosis of diabetes mellitus and a severe HP are present. Key Words: acute pancreatitis, diabetic ketoacidosis, hypertriglyceridemia Abbreviations: AP - acute pancreatitis, HP - hypertriglyceridemic pancreatitis, DKA - diabetic ketoacidosis (Pancreas 2015;44: 615–618)
H
ypertriglyceridemia is an uncommon cause of acute pancreatitis (AP) accounting for up to 10% of the cases. Usually, it occurs when serum triglyceride levels are greater than 1000 mg/dL.1–3 The coexistence of AP, diabetic ketoacidosis (DKA), and hypertriglyceridemia has been recognized previously in several case reports with a few patients, most of them related to poor diabetes control,2,4 alcoholism, pregnancy, prior AP, and personal or family history of hyperlipidemia.5 This association has also been related to uncommon conditions, such as acromegaly,6 and medication use, such as antidepressants, estrogens, furosemide, isotretinoin, tamoxifen, and β-blockers.1,7 Both AP and hypertriglyceridemia can be found as a cause and a consequence of DKA. Hypertriglyceridemic pancreatitis From the *Department of Internal Medicine, Dr José Eleuterio González University Hospital, †School of Medicine, ‡Endocrinology Division, Dr José Eleuterio González University Hospital, and §Research Division, School of Medicine, Universidad Autónoma de Nuevo León, Monterrey, Nuevo León, México. Received for publication February 27, 2014; accepted September 24, 2014. Reprints: Dania Lizet Quintanilla-Flores, MD, Department of Internal Medicine, Dr José Eleuterio González University Hospital, Universidad Autónoma de Nuevo León, Av. Francisco I. Madero pte. y Av. Gonzalitos s/n, Col. Mitras Centro, Monterrey, Nuevo León, México, 64460 (e‐mail: [email protected]). The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pancreas • Volume 44, Number 4, May 2015
(HP) occurs as a complication of DKA in at least 15% of the cases leading to a more severe episode of DKA with marked acidosis and hyperglycemia.8 Diabetic ketoacidosis, on the other hand, may complicate the clinical course of HP by several mechanisms; (1) because abdominal pain, hyperlipasemia, and hyperamylasemia can be found altogether in both conditions,9,10 the clinical diagnosis of AP is often delayed when DKA is the diagnosis on presentation; (2) the coexistence of DKA and AP both increase the intravascular volume depletion and disrupt the glucose homeostasis, which may complicate the clinical course of both conditions leading to increased morbidity and mortality rates.8 In addition, (3) hypertriglyceridemia in DKA may lead to AP, increasing the risk of both edematous and necrotizing pancreatitides.11 Given that both conditions may have similar clinical course, their diagnosis often represents a challenge to the attending physician. An early diagnosis of DKA in HP and vice versa is important to pursue an appropriated management and follow-up of these patients. It is believed that the presence of DKA has a direct impact in the clinical course of HP, leading to longer hospital say, a more severe disease, and a higher rate of complications and mortality compared with the consequences of HP alone; however, there is sparse information assessing this association. Our objective was to contrast the clinical and demographic differences in cases with AP, DKA, and hypertriglyceridemia compared with those patients without DKA and compare the severity of the illness, hospital stay, delay in oral intake, duration of insulin treatment, and several biochemical parameters.
MATERIALS AND METHODS Background We developed a retrospective, consecutive, observational study in the department of internal medicine at the Dr José Eleuterio González University Hospital from 2007 to 2011. Among 425 analyzed medical records with the diagnosis of AP, 62 adult patients were classified with HP. Eight patients with the coexistence of DKA were compared with all HP patients without DKA in a cross-sectional design. The study protocol was approved by the internal institutional review board.
Study Subjects Study subjects included all men and nonpregnant women aged older than 16 years with HP who completed the following inclusion criteria: complete medical record, diagnosis of AP according to the 2012 revision of the Atlanta classification and definitions of AP10 (abdominal pain, serum lipase or amylase activity at least 3 times greater than the upper limit of normal, and characteristic findings of AP on contrast-enhanced computer tomography), serum triglyceride levels greater than or equal to 1000 mg/dL, and abdomen ultrasound ruling out gallstones. Patients were excluded if the medical records were uncompleted or www.pancreasjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
615
Pancreas • Volume 44, Number 4, May 2015
Quintanilla-Flores et al
if the diagnosis of AP was doubtful according to the previous inclusion criteria. Among the 62 participants, 55 subjects (88%) had complete information for analysis. A total of 8 patients were diagnosed with DKA according to the American Diabetes Association recommendations12: serum glucose level greater than 250 mg/dL, paterial pH less than 7.30, serum bicarbonate less than 18 mEq/L, positive urine and serum ketones and anion gap greater than 10. There were 3 women (37.5%) and 5 men (62.5%), with a mean (SD) age of 32.5 (10.5) years. Non-DKA patients included 10 women (21.3%) and 37 men (78.7%), with a mean (SD) age of 39.2 (8.6) years.
Study Variables We analyzed age, sex, family history of lipid disorders, previous diagnosis of diabetes mellitus (DM), and hypertriglyceridemia as well as previous episodes of AP. Severity criteria were considered according to the Ranson score and APACHE II score.13 Blood samples were collected at admission for laboratory evaluation. We compared serum amylase and lipase values, leukocyte count, plasma glucose, lactic dehydrogenase, triglyceride levels, and Δ of change in triglycerides 24 hours after treatment initiation. In addition, we compared the hospital stay, mean insulin dose, hospital use of fibrates as well as omega 3, insulin dose at discharge, and mortality.
Data Analysis SPSS version 19 was used for statistical analysis. Data are expressed as mean (SD), median, minimum-maximum, or percentage. The differences in means of continuous measurements were compared using Student t test. The χ2 test was used for categorical data. Statistical significance was considered when P < 0.05.
RESULTS Global Study Population Table 1 shows the clinical findings of the study population. Fifty-five patients with HP were hospitalized during the study period, and an additional diagnosis of DKA was established at admission in 8 cases (14.5%). The mean (SD) age was 38.2 (9.1) years, and 13 subjects (23.6%) were female. A family history of dyslipidemia was detected in 18.1% of the participants. Personal history of DM type 2 was found in more than two thirds of the subjects, and almost half of the subjects had a previous established diagnosis of dyslipidemia (45.4%). Twenty-seven of the subjects had at least 1 previous episode of severe HP. Table 2 shows the main biochemical differences among patients with HP and without DKA. The Ranson and the APACHE II scores were statistically higher in the DKA group (P < 0.001). In addition, the levels of leukocyte count and serum glucose were more elevated in the DKA group compared with that in the nonDKA group (P = 0.07 and 0.001, respectively). There were no statistically significant differences between the remaining biochemical variables. Diabetic ketoacidosis patients were treated with higher insulin doses compared with non-DKA patients (P = 0.030). Hospital stay, mortality, delay in oral intake, and insulin dose at discharge were similar in both groups. The DKA group presented 2 major complications associated with AP: 1 patient with a pancreatic abscess and the other with acute respiratory distress syndrome. Diabetic ketoacidosis patients had increased rates of hospital readmission at 30 days of discharge compared with non-DKA patients (4 [50%] vs 5 [9.1%], P = 0.01).
DISCUSSION We describe the first case-control study that highlights the clinical course of patients with HP and the coexistence of DKA. The main features that we found included the following: (1) the
TABLE 1. Clinical Features of Patients in Both the DKA Group and the Control Group
n, % Age, mean (SD), y Female, n (%) Family history of lipid disorders, n (%) Previous episodes of pancreatitis, n (%) DM, n (%) Body mass index, mean (SD) Previous diagnosis of hypertriglyceridemia, n (%) Hospital stay, mean (SD), d Mean insulin dose, mean (SD), U/kg per h Delay in oral intake, mean (SD), d Duration of insulin infusion, mean (SD), d Fibrates, n (%) Omega 3, n (%) Insulin at discharge, n (%) Mean insulin dose at discharge, mean (SD), U/d Death, n (%) Hospitalization 30 d after discharge, n (%)
All Groups
DKA
Non-DKA Group
P
55 38.2 (9.1) 13 (23.6) 10 (18.1) 15 (27.2) 38 (69.0) 29.8 (4.0) 25 (45.4) 7.3 (3.5) 0.07 (0.08) 3.7 (1.4) 4.4 (2.6) 44 (80.0) 24 (43.6) 36 (65.5) 54.0 (39.1) 1 (1.8) 5 (9.1)
8 32.5 (10.5) 3 (37.5)
47 39.2 (8.6) 10 (21.3) 10 (21.3) 14 (29.7) 31 (65.9) 30.0 (4.0) 22 (46.8) 7 (3) 0.07 (0.08) 3.6 (1.3) 4.1 (2.3) 39 (82.9) 23 (48.9) 30 (63.8) 49.9 (39.6) 1 (2.1) 1 (2.1)
0.05 0.31 0.17 0.31 0.22 0.46 0.85 0.07 0.03 0.21 0.14 0.13 0.06 0.28 0.18 0.67 0.001
1 (12.5) 7 (87.5) 28.2 (4.0) 3 (37.5) 9 (4) 0.09 (0.04) 4.3 (1.9) 5.7 (3.9) 5 (62.5) 1 (12.5) 6 (75.0) 73.7 (32.2) 4 (50.0)
Group 1, DKA and HP; group 2, HP. d indicates days; h, hours; kg, kilogram; SD, standard deviation; U, units.
616
www.pancreasjournal.com
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pancreas • Volume 44, Number 4, May 2015
Hypertriglyceridemic Pancreatitis and Ketoacidosis
TABLE 2. Biochemical Features of Patients in Both the DKA Group and the Control Group All Groups
DKA
n 55 8 Ranson score, mean (SD) 1.9 (1.8) 3.6 (1.6) APACHE II score, mean (SD) 5.8 (3.9) 9.8 (3.6) Amylase, median (min-max), U/mL 490 (33-5154) 306 (114-1255) Lipase, median (min-max), (n = 16), U/mL 1021 (23-4296) 1200 (931-1200) 12.7 (5.7-34.4) 15.9 (11.8-25.0) Leukocyte count, median (min-max), 109/L Glucose, median (min-max), mg/dL 255 (84-739) 429 (306-739 Lactic dehydrogenase, median (min-max), UI/L 202 (61-779) 421 (61-642) Triglycerides, median (min-max), mg/dL 2918 (264-16168) 2276 (872-16168) Triglycerides 24 h after initiation of treatment, median (min-max), mg/dL 489 (190-2564) 383 (190-883) Triglycerides at discharge, median (min-max), mg/dL 348 (43-941) 359 (178-452)
Non-DKA Group
P
47 1.6 (1.7) 5.2 (3.6) 496 (33-5154) 950 (23-4296) 11.8 (5.7-34.4) 242 (84-505) 184 (86-779) 2918 (264-8040) 491 (196-2564) 348 (43-941)
0.003 0.005 0.62 0.98 0.07 0.001 0.16 0.08 0.43 0.75
Min-max, minimum-maximum.
DKA group presented higher serum glucose levels as well as more severe HP according to the Ranson score and the APACHE II score; (2) we found no differences in hospital stay, mortality, clinical course, or delay in oral intake in both groups; and (3) as expected, the DKA group was treated with higher doses of insulin infusion, although the duration of treatment as well as the clinical and biochemical response were similar. The association between DKA, AP, and hypertriglyceridemia has been reported in several case reports since 1980,2,4,6–8,14–17 concluding that hypertriglyceridemia in DKA represents a direct metabolic impact of the diabetic ketotic state and that transient but marked hypertryglyceridemia may lead to AP.15,18 In DKA, insulin deficiency activates lipolysis in the adipose tissue leading to the release of free fatty acids, which accelerates formation of very low-density lipoproteins in the liver. In addition, the reduced activity of lipoprotein lipase in peripheral tissue decreases removal of very low-density lipoproteins from the plasma, resulting in hypertriglyceridemia.2 The development of HP can be explained by several mechanisms: (1) the impairment of circulatory flow in capillary beds and the development of toxic injury to acinar cells and capillary endothelia by excessive free fatty acids, with the resulting pancreatic ischemia, (2) the subsequent release of inflammatory mediators and free radicals that lead to necrosis, edema, and inflammation; (3) the development of an hyperviscosity state that leads to ischemia and acidosis in pancreatic capillaries; and (4) the impaired triglyceride metabolism due to mutations in the lipoprotein lipase genes and dysregulation or deficiency of specific key enzymes and their substrates.1–3 Diabetic ketoacidosis is known to be an independent risk factor for HP as shown in a prospective study by Nair et al.8 On the other hand, younger age at diagnosis of hyperlipidemia and maximal triglyceride levels seem to represent risk factors for severe AP in patients with a history of hypertriglyceridemic disease.19 In our study, although not statistically significant, younger patients with an age of 35 years or younger seem to represent an increased risk for DKA in HP. In addition, a previous known history of DM and higher levels of serum glucose on admission should prompt the physicians to suspect the coexistence of DKA and AP. All these findings should be considered in the initial evaluation of any patient with DKA who fits in our population group. We found nonspecific elevations of serum amylase in 3 subjects in the DKA group and in 10 subjects in the control group. These results confirm previous reports that either normal or nonspecific serum amylase does not discard or correlate with the © 2015 Wolters Kluwer Health, Inc. All rights reserved.
development of AP in DKA.8,15 One considerable explanation is believed to be the interference with in vitro determination of amylase levels by disturbance of the calorimetric method as a consequence of hypertryglyceridemia and a lactescent serum.2 On the other hand, we found no statistical differences in serum amylase between both groups when it was found elevated, suggesting that other diagnostic tools than serum amylase should be considered when suspecting the coexistence of AP in DKA and hypertriglyceridemia. It is well known that by increased depletion of the intravascular volume, AP precipitates a more severe episode of DKA with marked acidosis and hyperglycemia, which in turn is associated with an increased mortality rate.8,20 On the other hand, the presence of DKA increases the severity of the AP represented by higher 48-hour Ranson score according to previous reports.2 Although DKA was associated with a higher 48-hour Ranson score and APACHE II score in our patients, the clinical course of the HP did not seem to differ in both groups evidenced by the similarities in the duration of hospital stay, duration of treatment with insulin infusion, mortality, and delay in oral intake. In addition, none of our patients developed statistically significant serious complications associated with the AP itself. It seems that the 48-hour Ranson score is likely to overestimate the clinical course of HP in patients with DKA, probably because of severe dehydration secondary to DKA, which increments the Ranson score despite a mild clinical course.8 The coexistence of DKA and HP should also prompt the physicians to promote a better discharge plan and frequent follow-up visits to avoid hospital readmission within the next 30 days, as evidenced in our findings, where patients in the DKA group had higher rates of hospitalization within the next 30 days after discharge. High prevalence of rehospitalization probably shows that AP was severe. Prolongation of hospitalizations would be considered to avoid early readmissions in patients with HP and DKA. The severity of HP does not correlate directly with the triglyceride levels, according to previous publications.5,19 Similarly, in our study, the presence of DKA did not affect the degree of hypertriglyceridemia on admission, evidenced by nonsignificant differences in serum triglyceride levels in both groups. In addition, instead of higher insulin dose in the DKA group, the Δ of triglyceride change at 24 hours of treatment was also incomparable. Triglyceride count should not be considered as a predictive factor of morbidity and mortality in HP. Since the 1990s, aggressive treatment with intravenous insulin infusion, along with fasting and heparin, has been mainstay www.pancreasjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
617
Pancreas • Volume 44, Number 4, May 2015
Quintanilla-Flores et al
in the treatment of HP to maintain euglycemia and normalize triglycerides rapidly.21 Our results show that although there were differences in insulin doses, the result in serum triglycerides was the same. The question arises whether aggressive treatment with insulin infusion directly influences triglyceride levels or if they can be treated just with fasting, low-dose insulin, and fluid replacement. Our study has several limitations inherent to an observational study; as a retrospective study, we cannot control all the variables studied or the treatment adjustments, which could bias our results. In addition, some of the factors that might be expected to predict higher risk of DKA in HP could not be measured, including anthropometric measures. As a major limitation, we identified the lack of power of this study and we recognize that larger and well-designed studies should be done. To our knowledge, this is the first study that evaluates the clinical course of AP in DKA compared with AP without DKA in hypertriglyceridemic patients. The coexistence of DKA and HP must be suspected when a severe AP is diagnosed. In addition, although DKA complicates the HP episode in terms of severity and leukocyte count, the clinical course during the hospital stay does not differ compared with HP without DKA. Our observations should be considered in future research studies and in the clinical evaluation of every patient with DKAwith abdominal pain and hypertriglyceridemia.
6. Lee CY, Lee MK, Lee SY, et al. A case of acromegaly with diabetic ketoacidosis and hypertriglyceridemia-induced acute pancreatitis. J Korean Soc Endocrinol. 2002;17:110–116. 7. Chen JL, Spinowitz N, Karwa M. Hypertriglyceridemia, acute pancreatitis, and diabetic ketoacidosis possibly associated with mirtazapine therapy: a case report. Pharmacotherapy. 2003;23:940–944. 8. Nair S, Yadav D, Pitchumoni CS. Association of diabetic ketoacidosis and acute pancreatitis: observations in 100 consecutive episodes of DKA. Am J Gastroenterol. 2000;95:2795–2800. 9. Knight AH, Williams DN, Ellis G, et al. Significance of hyperamylasemia and abdominal pain in diabetic ketoacidosis. BMJ. 1973;3:128–131. 10. Banks PA, Bollen TL, Dervenis C, et al. Classification of acute pancreatitis—2012: revision of the Atlanta classification and definitions by international consensus. Gut. 2013;62:102–111. 11. Deng L, Xue P, Xia Q, et al. Effect of admission hypertriglyceridemia on the episodes of severe acute pancreatitis. World J Gastroenterol. 2008;14: 4558–4561. 12. Kitabchi AE, Umpierrez GE, Miles JM, et al. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009;32:1335–1343. 13. Banks BA, Freeman ML, Fass DS, et al. Practice guidelines in acute pancreatitis. Am J Gastroenterol. 2006;101:2379–2400. 14. Winter RJ, Herr TJ, Stoner NJ, et al. Diabetic lipemia in childhood diabetic ketoacidosis: a clue to coexisting acute pancreatitis. Diabetes Care. 1980;3: 706–707.
ACKNOWLEDGMENT The authors wish to thank Dr Sergio Lozano for the English translation. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
15. Nair S, Pitchumoni CS. Diabetic ketoacidosis, hyperlipidemia, and acute pancreatitis: the enigmatic triangle. Am J Gastroenterol. 1997;92: 1560–1561.
REFERENCES
16. Denecker N, Decochez K. Poorly controlled type 2 diabetes complicated by an episode of severe hypertriglyceridaemia-induced pancreatitis. BMJ Case Rep. 2013. doi: 10.1136/bcr-2012-008455.
1. Gan SI, Edwards AL, Symonds CJ, et al. Hypertriglyceridemia-induced pancreatitis: a case-based review. World J Gastroenterol. 2006;12: 7197–7202. 2. Hahn SJ, Park JH, Lee JH, et al. Severe hypertriglyceridemia in diabetic ketoacidosis accompanied by acute pancreatitis: case report. J Korean Med Sci. 2010;25:1375–1378.
17. Shu S, Yoshikazu U, Takahiro F, et al. A case of diabetic ketoacidosis complicated by severe hypertriglyceridemia and acute pancreatitis. J Japan Diab Soc. 2000;43:561–566. 18. Gianfrate L, Ferraris L. Acute pancreatitis, hyperlipidemia, and diabetic ketoacidosis: who comes first? Am J Gastroenterol. 1998;93:1393–1394.
3. Ewald N, Hardt PD, Kloer H-U. Severe hypertriglyceridemia and pancreatitis: presentation and management. Curr Opin Lipidol. 2009;20: 497–504.
19. Lloret-Linares C, Pelletier AL, Czernichow S, et al. Acute pancreatitis in a cohort of 129 patients referred for severe hypertriglyceridemia. Pancreas. 2008;37:13–2.
4. Ogawa D, Wada J, Makino H. Diabetic lipemia associated with acute pancreatitis in a patient with type 2 diabetes. J Diabetes Mellitus. 2011;1: 54–56.
20. Nielson OS, Simonsen E. A case of transient diabetes mellitus in connection with acute pancreatitis. Acta Med Scand. 1969;185:459–462.
5. Tsuang W, Navaneethan U, Ruiz L, et al. Hypertrigliceridemic pancreatitis: presentation and management. Am J Gastroenterol. 2009;104:984–991.
618
www.pancreasjournal.com
21. Hensen C, Rock M, Schnieper C, et al. Heparin and insulin in the treatment of acute hypertriglyceridemia-induced pancreatitis [in German]. Schweiz Med Wochenschr. 1999;129:1242–1248.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Clinical Course of Diabetic Ketoacidosis in Hypertriglyceridemic Pancreatitis Dania Lizet Quintanilla-Flores, MD,* Erick Joel Rendón-Ramírez, MD,* Perla Rocío Colunga-Pedraza, MD,* Jesús Gallardo-Escamilla, MD,† Sergio Antonio Corral-Benavides, MD,† José Gerardo González-González, PhD,‡§ and Héctor Eloy Tamez-Pérez, PhD§ Objectives: Hypertriglyceridemic pancreatitis (HP) is an uncommon condition accounting for 1% to 4% of cases of acute pancreatitis, mostly associated with poor glycemic control. Diabetic ketoacidosis (DKA) may complicate the clinical course of HP. Our objective was to identify clinical and demographic differences between HP and DKA patients compared with those without DKA. Methods: Fifty-five patients with HP were included. Diabetic ketoacidosis was diagnosed in 8 patients. We analyzed the severity, hospital stay, delay in oral intake, duration of insulin infusion, complete blood cell count, and triglyceride levels. Results: Diabetic ketoacidosis was associated with a more severe HP. There were no differences in hospital stay, delay in oral intake, or duration of insulin treatment in both groups. Serum amylase, lipase, and triglyceride levels were similar. Previous diagnosis of diabetes mellitus, higher Ranson and APACHE II scores, and higher serum glucose level at admission were the only predictive risk factors for DKA and HP. Conclusions: Coexistence of DKA does not modify the clinical course of HP, although a more severe episode of HP in DKA patients. Diabetic ketoacidosis was associated with higher insulin doses, without impact in triglyceride levels. Diabetic ketoacidosis and HP should be considered when a previous diagnosis of diabetes mellitus and a severe HP are present. Key Words: acute pancreatitis, diabetic ketoacidosis, hypertriglyceridemia Abbreviations: AP - acute pancreatitis, HP - hypertriglyceridemic pancreatitis, DKA - diabetic ketoacidosis (Pancreas 2015;44: 615–618)
H
ypertriglyceridemia is an uncommon cause of acute pancreatitis (AP) accounting for up to 10% of the cases. Usually, it occurs when serum triglyceride levels are greater than 1000 mg/dL.1–3 The coexistence of AP, diabetic ketoacidosis (DKA), and hypertriglyceridemia has been recognized previously in several case reports with a few patients, most of them related to poor diabetes control,2,4 alcoholism, pregnancy, prior AP, and personal or family history of hyperlipidemia.5 This association has also been related to uncommon conditions, such as acromegaly,6 and medication use, such as antidepressants, estrogens, furosemide, isotretinoin, tamoxifen, and β-blockers.1,7 Both AP and hypertriglyceridemia can be found as a cause and a consequence of DKA. Hypertriglyceridemic pancreatitis From the *Department of Internal Medicine, Dr José Eleuterio González University Hospital, †School of Medicine, ‡Endocrinology Division, Dr José Eleuterio González University Hospital, and §Research Division, School of Medicine, Universidad Autónoma de Nuevo León, Monterrey, Nuevo León, México. Received for publication February 27, 2014; accepted September 24, 2014. Reprints: Dania Lizet Quintanilla-Flores, MD, Department of Internal Medicine, Dr José Eleuterio González University Hospital, Universidad Autónoma de Nuevo León, Av. Francisco I. Madero pte. y Av. Gonzalitos s/n, Col. Mitras Centro, Monterrey, Nuevo León, México, 64460 (e‐mail: [email protected]). The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pancreas • Volume 44, Number 4, May 2015
(HP) occurs as a complication of DKA in at least 15% of the cases leading to a more severe episode of DKA with marked acidosis and hyperglycemia.8 Diabetic ketoacidosis, on the other hand, may complicate the clinical course of HP by several mechanisms; (1) because abdominal pain, hyperlipasemia, and hyperamylasemia can be found altogether in both conditions,9,10 the clinical diagnosis of AP is often delayed when DKA is the diagnosis on presentation; (2) the coexistence of DKA and AP both increase the intravascular volume depletion and disrupt the glucose homeostasis, which may complicate the clinical course of both conditions leading to increased morbidity and mortality rates.8 In addition, (3) hypertriglyceridemia in DKA may lead to AP, increasing the risk of both edematous and necrotizing pancreatitides.11 Given that both conditions may have similar clinical course, their diagnosis often represents a challenge to the attending physician. An early diagnosis of DKA in HP and vice versa is important to pursue an appropriated management and follow-up of these patients. It is believed that the presence of DKA has a direct impact in the clinical course of HP, leading to longer hospital say, a more severe disease, and a higher rate of complications and mortality compared with the consequences of HP alone; however, there is sparse information assessing this association. Our objective was to contrast the clinical and demographic differences in cases with AP, DKA, and hypertriglyceridemia compared with those patients without DKA and compare the severity of the illness, hospital stay, delay in oral intake, duration of insulin treatment, and several biochemical parameters.
MATERIALS AND METHODS Background We developed a retrospective, consecutive, observational study in the department of internal medicine at the Dr José Eleuterio González University Hospital from 2007 to 2011. Among 425 analyzed medical records with the diagnosis of AP, 62 adult patients were classified with HP. Eight patients with the coexistence of DKA were compared with all HP patients without DKA in a cross-sectional design. The study protocol was approved by the internal institutional review board.
Study Subjects Study subjects included all men and nonpregnant women aged older than 16 years with HP who completed the following inclusion criteria: complete medical record, diagnosis of AP according to the 2012 revision of the Atlanta classification and definitions of AP10 (abdominal pain, serum lipase or amylase activity at least 3 times greater than the upper limit of normal, and characteristic findings of AP on contrast-enhanced computer tomography), serum triglyceride levels greater than or equal to 1000 mg/dL, and abdomen ultrasound ruling out gallstones. Patients were excluded if the medical records were uncompleted or www.pancreasjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
615
Pancreas • Volume 44, Number 4, May 2015
Quintanilla-Flores et al
if the diagnosis of AP was doubtful according to the previous inclusion criteria. Among the 62 participants, 55 subjects (88%) had complete information for analysis. A total of 8 patients were diagnosed with DKA according to the American Diabetes Association recommendations12: serum glucose level greater than 250 mg/dL, paterial pH less than 7.30, serum bicarbonate less than 18 mEq/L, positive urine and serum ketones and anion gap greater than 10. There were 3 women (37.5%) and 5 men (62.5%), with a mean (SD) age of 32.5 (10.5) years. Non-DKA patients included 10 women (21.3%) and 37 men (78.7%), with a mean (SD) age of 39.2 (8.6) years.
Study Variables We analyzed age, sex, family history of lipid disorders, previous diagnosis of diabetes mellitus (DM), and hypertriglyceridemia as well as previous episodes of AP. Severity criteria were considered according to the Ranson score and APACHE II score.13 Blood samples were collected at admission for laboratory evaluation. We compared serum amylase and lipase values, leukocyte count, plasma glucose, lactic dehydrogenase, triglyceride levels, and Δ of change in triglycerides 24 hours after treatment initiation. In addition, we compared the hospital stay, mean insulin dose, hospital use of fibrates as well as omega 3, insulin dose at discharge, and mortality.
Data Analysis SPSS version 19 was used for statistical analysis. Data are expressed as mean (SD), median, minimum-maximum, or percentage. The differences in means of continuous measurements were compared using Student t test. The χ2 test was used for categorical data. Statistical significance was considered when P < 0.05.
RESULTS Global Study Population Table 1 shows the clinical findings of the study population. Fifty-five patients with HP were hospitalized during the study period, and an additional diagnosis of DKA was established at admission in 8 cases (14.5%). The mean (SD) age was 38.2 (9.1) years, and 13 subjects (23.6%) were female. A family history of dyslipidemia was detected in 18.1% of the participants. Personal history of DM type 2 was found in more than two thirds of the subjects, and almost half of the subjects had a previous established diagnosis of dyslipidemia (45.4%). Twenty-seven of the subjects had at least 1 previous episode of severe HP. Table 2 shows the main biochemical differences among patients with HP and without DKA. The Ranson and the APACHE II scores were statistically higher in the DKA group (P < 0.001). In addition, the levels of leukocyte count and serum glucose were more elevated in the DKA group compared with that in the nonDKA group (P = 0.07 and 0.001, respectively). There were no statistically significant differences between the remaining biochemical variables. Diabetic ketoacidosis patients were treated with higher insulin doses compared with non-DKA patients (P = 0.030). Hospital stay, mortality, delay in oral intake, and insulin dose at discharge were similar in both groups. The DKA group presented 2 major complications associated with AP: 1 patient with a pancreatic abscess and the other with acute respiratory distress syndrome. Diabetic ketoacidosis patients had increased rates of hospital readmission at 30 days of discharge compared with non-DKA patients (4 [50%] vs 5 [9.1%], P = 0.01).
DISCUSSION We describe the first case-control study that highlights the clinical course of patients with HP and the coexistence of DKA. The main features that we found included the following: (1) the
TABLE 1. Clinical Features of Patients in Both the DKA Group and the Control Group
n, % Age, mean (SD), y Female, n (%) Family history of lipid disorders, n (%) Previous episodes of pancreatitis, n (%) DM, n (%) Body mass index, mean (SD) Previous diagnosis of hypertriglyceridemia, n (%) Hospital stay, mean (SD), d Mean insulin dose, mean (SD), U/kg per h Delay in oral intake, mean (SD), d Duration of insulin infusion, mean (SD), d Fibrates, n (%) Omega 3, n (%) Insulin at discharge, n (%) Mean insulin dose at discharge, mean (SD), U/d Death, n (%) Hospitalization 30 d after discharge, n (%)
All Groups
DKA
Non-DKA Group
P
55 38.2 (9.1) 13 (23.6) 10 (18.1) 15 (27.2) 38 (69.0) 29.8 (4.0) 25 (45.4) 7.3 (3.5) 0.07 (0.08) 3.7 (1.4) 4.4 (2.6) 44 (80.0) 24 (43.6) 36 (65.5) 54.0 (39.1) 1 (1.8) 5 (9.1)
8 32.5 (10.5) 3 (37.5)
47 39.2 (8.6) 10 (21.3) 10 (21.3) 14 (29.7) 31 (65.9) 30.0 (4.0) 22 (46.8) 7 (3) 0.07 (0.08) 3.6 (1.3) 4.1 (2.3) 39 (82.9) 23 (48.9) 30 (63.8) 49.9 (39.6) 1 (2.1) 1 (2.1)
0.05 0.31 0.17 0.31 0.22 0.46 0.85 0.07 0.03 0.21 0.14 0.13 0.06 0.28 0.18 0.67 0.001
1 (12.5) 7 (87.5) 28.2 (4.0) 3 (37.5) 9 (4) 0.09 (0.04) 4.3 (1.9) 5.7 (3.9) 5 (62.5) 1 (12.5) 6 (75.0) 73.7 (32.2) 4 (50.0)
Group 1, DKA and HP; group 2, HP. d indicates days; h, hours; kg, kilogram; SD, standard deviation; U, units.
616
www.pancreasjournal.com
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pancreas • Volume 44, Number 4, May 2015
Hypertriglyceridemic Pancreatitis and Ketoacidosis
TABLE 2. Biochemical Features of Patients in Both the DKA Group and the Control Group All Groups
DKA
n 55 8 Ranson score, mean (SD) 1.9 (1.8) 3.6 (1.6) APACHE II score, mean (SD) 5.8 (3.9) 9.8 (3.6) Amylase, median (min-max), U/mL 490 (33-5154) 306 (114-1255) Lipase, median (min-max), (n = 16), U/mL 1021 (23-4296) 1200 (931-1200) 12.7 (5.7-34.4) 15.9 (11.8-25.0) Leukocyte count, median (min-max), 109/L Glucose, median (min-max), mg/dL 255 (84-739) 429 (306-739 Lactic dehydrogenase, median (min-max), UI/L 202 (61-779) 421 (61-642) Triglycerides, median (min-max), mg/dL 2918 (264-16168) 2276 (872-16168) Triglycerides 24 h after initiation of treatment, median (min-max), mg/dL 489 (190-2564) 383 (190-883) Triglycerides at discharge, median (min-max), mg/dL 348 (43-941) 359 (178-452)
Non-DKA Group
P
47 1.6 (1.7) 5.2 (3.6) 496 (33-5154) 950 (23-4296) 11.8 (5.7-34.4) 242 (84-505) 184 (86-779) 2918 (264-8040) 491 (196-2564) 348 (43-941)
0.003 0.005 0.62 0.98 0.07 0.001 0.16 0.08 0.43 0.75
Min-max, minimum-maximum.
DKA group presented higher serum glucose levels as well as more severe HP according to the Ranson score and the APACHE II score; (2) we found no differences in hospital stay, mortality, clinical course, or delay in oral intake in both groups; and (3) as expected, the DKA group was treated with higher doses of insulin infusion, although the duration of treatment as well as the clinical and biochemical response were similar. The association between DKA, AP, and hypertriglyceridemia has been reported in several case reports since 1980,2,4,6–8,14–17 concluding that hypertriglyceridemia in DKA represents a direct metabolic impact of the diabetic ketotic state and that transient but marked hypertryglyceridemia may lead to AP.15,18 In DKA, insulin deficiency activates lipolysis in the adipose tissue leading to the release of free fatty acids, which accelerates formation of very low-density lipoproteins in the liver. In addition, the reduced activity of lipoprotein lipase in peripheral tissue decreases removal of very low-density lipoproteins from the plasma, resulting in hypertriglyceridemia.2 The development of HP can be explained by several mechanisms: (1) the impairment of circulatory flow in capillary beds and the development of toxic injury to acinar cells and capillary endothelia by excessive free fatty acids, with the resulting pancreatic ischemia, (2) the subsequent release of inflammatory mediators and free radicals that lead to necrosis, edema, and inflammation; (3) the development of an hyperviscosity state that leads to ischemia and acidosis in pancreatic capillaries; and (4) the impaired triglyceride metabolism due to mutations in the lipoprotein lipase genes and dysregulation or deficiency of specific key enzymes and their substrates.1–3 Diabetic ketoacidosis is known to be an independent risk factor for HP as shown in a prospective study by Nair et al.8 On the other hand, younger age at diagnosis of hyperlipidemia and maximal triglyceride levels seem to represent risk factors for severe AP in patients with a history of hypertriglyceridemic disease.19 In our study, although not statistically significant, younger patients with an age of 35 years or younger seem to represent an increased risk for DKA in HP. In addition, a previous known history of DM and higher levels of serum glucose on admission should prompt the physicians to suspect the coexistence of DKA and AP. All these findings should be considered in the initial evaluation of any patient with DKA who fits in our population group. We found nonspecific elevations of serum amylase in 3 subjects in the DKA group and in 10 subjects in the control group. These results confirm previous reports that either normal or nonspecific serum amylase does not discard or correlate with the © 2015 Wolters Kluwer Health, Inc. All rights reserved.
development of AP in DKA.8,15 One considerable explanation is believed to be the interference with in vitro determination of amylase levels by disturbance of the calorimetric method as a consequence of hypertryglyceridemia and a lactescent serum.2 On the other hand, we found no statistical differences in serum amylase between both groups when it was found elevated, suggesting that other diagnostic tools than serum amylase should be considered when suspecting the coexistence of AP in DKA and hypertriglyceridemia. It is well known that by increased depletion of the intravascular volume, AP precipitates a more severe episode of DKA with marked acidosis and hyperglycemia, which in turn is associated with an increased mortality rate.8,20 On the other hand, the presence of DKA increases the severity of the AP represented by higher 48-hour Ranson score according to previous reports.2 Although DKA was associated with a higher 48-hour Ranson score and APACHE II score in our patients, the clinical course of the HP did not seem to differ in both groups evidenced by the similarities in the duration of hospital stay, duration of treatment with insulin infusion, mortality, and delay in oral intake. In addition, none of our patients developed statistically significant serious complications associated with the AP itself. It seems that the 48-hour Ranson score is likely to overestimate the clinical course of HP in patients with DKA, probably because of severe dehydration secondary to DKA, which increments the Ranson score despite a mild clinical course.8 The coexistence of DKA and HP should also prompt the physicians to promote a better discharge plan and frequent follow-up visits to avoid hospital readmission within the next 30 days, as evidenced in our findings, where patients in the DKA group had higher rates of hospitalization within the next 30 days after discharge. High prevalence of rehospitalization probably shows that AP was severe. Prolongation of hospitalizations would be considered to avoid early readmissions in patients with HP and DKA. The severity of HP does not correlate directly with the triglyceride levels, according to previous publications.5,19 Similarly, in our study, the presence of DKA did not affect the degree of hypertriglyceridemia on admission, evidenced by nonsignificant differences in serum triglyceride levels in both groups. In addition, instead of higher insulin dose in the DKA group, the Δ of triglyceride change at 24 hours of treatment was also incomparable. Triglyceride count should not be considered as a predictive factor of morbidity and mortality in HP. Since the 1990s, aggressive treatment with intravenous insulin infusion, along with fasting and heparin, has been mainstay www.pancreasjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
617
Pancreas • Volume 44, Number 4, May 2015
Quintanilla-Flores et al
in the treatment of HP to maintain euglycemia and normalize triglycerides rapidly.21 Our results show that although there were differences in insulin doses, the result in serum triglycerides was the same. The question arises whether aggressive treatment with insulin infusion directly influences triglyceride levels or if they can be treated just with fasting, low-dose insulin, and fluid replacement. Our study has several limitations inherent to an observational study; as a retrospective study, we cannot control all the variables studied or the treatment adjustments, which could bias our results. In addition, some of the factors that might be expected to predict higher risk of DKA in HP could not be measured, including anthropometric measures. As a major limitation, we identified the lack of power of this study and we recognize that larger and well-designed studies should be done. To our knowledge, this is the first study that evaluates the clinical course of AP in DKA compared with AP without DKA in hypertriglyceridemic patients. The coexistence of DKA and HP must be suspected when a severe AP is diagnosed. In addition, although DKA complicates the HP episode in terms of severity and leukocyte count, the clinical course during the hospital stay does not differ compared with HP without DKA. Our observations should be considered in future research studies and in the clinical evaluation of every patient with DKAwith abdominal pain and hypertriglyceridemia.
6. Lee CY, Lee MK, Lee SY, et al. A case of acromegaly with diabetic ketoacidosis and hypertriglyceridemia-induced acute pancreatitis. J Korean Soc Endocrinol. 2002;17:110–116. 7. Chen JL, Spinowitz N, Karwa M. Hypertriglyceridemia, acute pancreatitis, and diabetic ketoacidosis possibly associated with mirtazapine therapy: a case report. Pharmacotherapy. 2003;23:940–944. 8. Nair S, Yadav D, Pitchumoni CS. Association of diabetic ketoacidosis and acute pancreatitis: observations in 100 consecutive episodes of DKA. Am J Gastroenterol. 2000;95:2795–2800. 9. Knight AH, Williams DN, Ellis G, et al. Significance of hyperamylasemia and abdominal pain in diabetic ketoacidosis. BMJ. 1973;3:128–131. 10. Banks PA, Bollen TL, Dervenis C, et al. Classification of acute pancreatitis—2012: revision of the Atlanta classification and definitions by international consensus. Gut. 2013;62:102–111. 11. Deng L, Xue P, Xia Q, et al. Effect of admission hypertriglyceridemia on the episodes of severe acute pancreatitis. World J Gastroenterol. 2008;14: 4558–4561. 12. Kitabchi AE, Umpierrez GE, Miles JM, et al. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009;32:1335–1343. 13. Banks BA, Freeman ML, Fass DS, et al. Practice guidelines in acute pancreatitis. Am J Gastroenterol. 2006;101:2379–2400. 14. Winter RJ, Herr TJ, Stoner NJ, et al. Diabetic lipemia in childhood diabetic ketoacidosis: a clue to coexisting acute pancreatitis. Diabetes Care. 1980;3: 706–707.
ACKNOWLEDGMENT The authors wish to thank Dr Sergio Lozano for the English translation. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
15. Nair S, Pitchumoni CS. Diabetic ketoacidosis, hyperlipidemia, and acute pancreatitis: the enigmatic triangle. Am J Gastroenterol. 1997;92: 1560–1561.
REFERENCES
16. Denecker N, Decochez K. Poorly controlled type 2 diabetes complicated by an episode of severe hypertriglyceridaemia-induced pancreatitis. BMJ Case Rep. 2013. doi: 10.1136/bcr-2012-008455.
1. Gan SI, Edwards AL, Symonds CJ, et al. Hypertriglyceridemia-induced pancreatitis: a case-based review. World J Gastroenterol. 2006;12: 7197–7202. 2. Hahn SJ, Park JH, Lee JH, et al. Severe hypertriglyceridemia in diabetic ketoacidosis accompanied by acute pancreatitis: case report. J Korean Med Sci. 2010;25:1375–1378.
17. Shu S, Yoshikazu U, Takahiro F, et al. A case of diabetic ketoacidosis complicated by severe hypertriglyceridemia and acute pancreatitis. J Japan Diab Soc. 2000;43:561–566. 18. Gianfrate L, Ferraris L. Acute pancreatitis, hyperlipidemia, and diabetic ketoacidosis: who comes first? Am J Gastroenterol. 1998;93:1393–1394.
3. Ewald N, Hardt PD, Kloer H-U. Severe hypertriglyceridemia and pancreatitis: presentation and management. Curr Opin Lipidol. 2009;20: 497–504.
19. Lloret-Linares C, Pelletier AL, Czernichow S, et al. Acute pancreatitis in a cohort of 129 patients referred for severe hypertriglyceridemia. Pancreas. 2008;37:13–2.
4. Ogawa D, Wada J, Makino H. Diabetic lipemia associated with acute pancreatitis in a patient with type 2 diabetes. J Diabetes Mellitus. 2011;1: 54–56.
20. Nielson OS, Simonsen E. A case of transient diabetes mellitus in connection with acute pancreatitis. Acta Med Scand. 1969;185:459–462.
5. Tsuang W, Navaneethan U, Ruiz L, et al. Hypertrigliceridemic pancreatitis: presentation and management. Am J Gastroenterol. 2009;104:984–991.
618
www.pancreasjournal.com
21. Hensen C, Rock M, Schnieper C, et al. Heparin and insulin in the treatment of acute hypertriglyceridemia-induced pancreatitis [in German]. Schweiz Med Wochenschr. 1999;129:1242–1248.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
