Newly Diagnosed Diabetes Mellitus After Acute Pancreatitis
Newly Diagnosed Diabetes Mellitus After Acute Pancreatitis
The search strategy was developed to identify all prospective clinical studies that reported the change in glucose homeostasis after discharge from hospital following the first episode of AP. An electronic literature search of three databases (MEDLINE, Scopus, and EMBASE) was conducted until 31 May 2013. The earliest available abstracts reviewed were from 1946 for MEDLINE, 1966 for Scopus and 1980 for EMBASE.
The search strategy used was as follows:
MEDLINE and EMBASE: 'acute pancreatitis' AND ('endocrine insufficiency' OR ' endocrine function' OR 'pancreatitis function' OR 'diabetes mellitus' OR 'pre diabetic state' OR 'type 2 diabetes mellitus' OR 'type 1 diabetes mellitus' OR 'adult onset diabetes mellitus' OR 'maturity onset diabetes' OR 'non-insulin dependent diabetes' OR 'insulin dependent diabetes' OR 'glucose intolerance' OR 'glucose homeostasis').
Scopus: 'acute pancreatitis' AND ('endocrine insufficiency' OR ' endocrine function' OR 'pancreatitis function' OR ' diabetes mellitus' OR 'pre diabetic state' OR 'type 2 diabetes mellitus' OR 'type 1 diabetes mellitus' OR 'adult onset diabetes mellitus' OR 'maturity onset diabetes' OR 'non-insulin dependent diabetes' OR 'insulin dependent diabetes' OR 'glucose intolerance' OR 'glucose homeostasis') AND (LIMIT-TO(EXACTKEYWORD, 'Human'))
Bibliographies of all included studies were also reviewed for other relevant articles. The search was limited to publications in English.
Only prospective clinical studies of adult patients (aged >18 years) with a first attack of AP and with a follow-up period of at least 1 month after hospital discharge were included. The main outcome measures studied were prediabetes, DM and DM treated with insulin.
Studies were excluded if they did not state that patients had no previous history of DM or prediabetes, or they did not provide data specifically on patients with no previous history of DM or prediabetes. Studies were also excluded for the following reasons: duplicate publication, retrospective study design, case report, did not use the standard definition of pancreatitis (the presence of two of the following three features: abdominal pain characteristic of AP, serum amylase and/or lipase three times the upper limit of normal and/or characteristic findings of AP on computed tomography scan), study population limited to patients who underwent pancreatic surgery such as necrosectomy or pancreatic resection, study population limited to patients with chronic, autoimmune, or hereditary pancreatitis, gestational diabetes.
Records extracted by the initial search were screened, and potentially relevant papers were retrieved and examined in more detail. Eligibility assessment was performed independently by two authors (SLMD and PPS) and any disagreement over study inclusion or exclusion was resolved by discussion with the senior author (MSP).
Data were independently extracted by two authors (SLMD and PPS) using a standardised data collection form. For each eligible study the following data were extracted: (1) year conducted, (2) country/countries, (3) number of participating centres, (4) study design (eg, cohort, case-control study), (5) source of control (if applicable), (6) method of data collection, (7) criteria used to diagnose AP and classify its severity, (8) number of individuals with complete data, (9) age, (10) proportion of male individuals, (11) Body Mass Index (BMI) of these individuals, (12) aetiology of AP (ie, biliary, alcohol, other), (13) duration of follow-up period, (14) glucose homeostasis test employed, (15) criteria used to define prediabetes and DM, (16) occurrence of prediabetes, DM and DM treated with insulin, (17) number of patients with pancreatic necrosis, and (18) the presence and influence on glucose homeostasis of further attacks of AP during follow-up period. If any of these parameters were not available from the published study, the corresponding authors were contacted to provide further information.
The Newcastle-Ottawa Scale (NOS), a widely accepted tool to assess the quality of observational studies, was used. It allocates a maximum of nine points based on three aspects of study design: patient selection, comparability of study groups, and exposure and outcome of study participants. Studies were considered to be of high quality if scored 5 or greater points, and of low quality if points scored were 4 and below.
Prediabetes was defined by specific fasting blood glucose (FBG) and/or 2 h oral glucose tolerance test (OGTT) criteria as reported by the study authors (FBG ≥5.6 mmol/L (100 mg/dL) and <7.0 mmol/L (126 mg/dL), or ≥ 6.1 mmol/L (110 mg/dL) and <7.0 mmol/L (126 mg/dL) and 2 h OGTT ≥7.8 mmol/L (140 mg/dL) and <11.1 (200 mg/dL)). DM was defined as reported by the authors of primary studies (FBG ≥7.8 mmol/L (140 mg/dL) or 2 h OGTT ≥11.1 mmol/L (200 mg/dL) or FBG ≥7.0 mmol/L (126 mg/dL)) and/or treatment with insulin, oral hypoglycaemia agents or specific dietary management. Treatment with insulin was defined as DM requiring permanent insulin therapy at follow-up (provided the mean follow-up time was equal to or greater than 3 months). Prevalence of prediabetes, DM and DM treated with insulin after AP was defined as the proportion of patients with each outcome listed after AP which was previously undiagnosed. Ideally we would have liked to report the cumulative incidence of prediabetes, DM and DM treated with insulin after AP, but given that regular universal screening for diabetes is uncommon and variable, we could not exclude pre-existing DM among those without history of diagnosed DM or prediabetes. We therefore chose to present data as prevalence of newly diagnosed DM over time since first attack of AP.
The 'time course' is defined at the prevalence of each outcome listed within certain time periods. This arbitrary includes 'up to 12 months', '12–36 month', '36–60 months' and 'more than 60 months'.
The definition of DM has changed over the years, and this is reflected in the different definitions used in the studies included in this review. The first widely accepted classification of DM was published by National Diabetes Data Group (fasting ≥7.8 h or 2 h ≥11.1 mmol/L) in 1979, WHO (fasting ≥8.0 h or 2 h ≥11.0 mmol/L) in 1980 and, in modified form, in 1985 (fasting ≥7.8 h or 2 h ≥11.1), and the next major revision occurred in 1997 by American Diabetes Association (ADA), in 1999 by WHO, and again in 2003 by ADA. The major change in the 1999 classification was the lowering of the diagnostic threshold of the fasting plasma glucose concentration. According to the 1999 WHO classification, DM was defined as a FBG of ≥7.0 mmol L (126 mg dL) or 2 h OGTT 2 h ≥11.1 mmol/L (200 mg/dL).
Studies were considered prospective if follow-up assessment was done prospectively. A binary classification was used to determine the severity of AP since the recent international multidisciplinary classification of AP severity, published by the Pancreatitis Across Nations Clinical Research and Education Alliance (PANCREA), was not used in the primary studies. This means that the study population was labelled as 'severe' if the studies reported on haemorrhagic AP, necrotising AP, or severe AP based on the Atlanta criteria. The remaining study populations were labelled as 'mild'.
All the eligible studies were combined to give a pooled prevalence for each of the outcomes studied. If the papers reported groups of individuals with different mean follow-up times, the group with the longest follow-up was analysed for pooled prevalence. Data were pooled using a random-effects model to give the most conservative estimate. Stats Direct V.2.7.9 (StatsDirect, 2008) was used to generate forest plots of pooled prevalence with 95% CIs. Statistical heterogeneity between the studies was assessed using the I metric with cut-offs of 25, 50 and 75% to define low, moderate and high heterogeneity, respectively. Publication bias was assessed by applying the Horbold–Egger's test to funnel plots. Subgroup analysis was conducted according to mean follow-up time in the study, and they were grouped as follows: up to 12 months, between 12 and 36 months, between 36 and 60 months, more than 60 months. Sensitivity analyses constrained to studies that employed the 1999 WHO definition of DM, and to studies with severe AP, were conducted. Metaregression was used to investigate whether sex, age, duration of follow-up and aetiology of AP (proportion of biliary or alcohol AP) had a significant influence on the summary estimates. The R 2.15.2 framework with package metafor was used for metaregression analysis. Reporting of this study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines.
Methods
Search Strategy and Study Selection
The search strategy was developed to identify all prospective clinical studies that reported the change in glucose homeostasis after discharge from hospital following the first episode of AP. An electronic literature search of three databases (MEDLINE, Scopus, and EMBASE) was conducted until 31 May 2013. The earliest available abstracts reviewed were from 1946 for MEDLINE, 1966 for Scopus and 1980 for EMBASE.
The search strategy used was as follows:
MEDLINE and EMBASE: 'acute pancreatitis' AND ('endocrine insufficiency' OR ' endocrine function' OR 'pancreatitis function' OR 'diabetes mellitus' OR 'pre diabetic state' OR 'type 2 diabetes mellitus' OR 'type 1 diabetes mellitus' OR 'adult onset diabetes mellitus' OR 'maturity onset diabetes' OR 'non-insulin dependent diabetes' OR 'insulin dependent diabetes' OR 'glucose intolerance' OR 'glucose homeostasis').
Scopus: 'acute pancreatitis' AND ('endocrine insufficiency' OR ' endocrine function' OR 'pancreatitis function' OR ' diabetes mellitus' OR 'pre diabetic state' OR 'type 2 diabetes mellitus' OR 'type 1 diabetes mellitus' OR 'adult onset diabetes mellitus' OR 'maturity onset diabetes' OR 'non-insulin dependent diabetes' OR 'insulin dependent diabetes' OR 'glucose intolerance' OR 'glucose homeostasis') AND (LIMIT-TO(EXACTKEYWORD, 'Human'))
Bibliographies of all included studies were also reviewed for other relevant articles. The search was limited to publications in English.
Only prospective clinical studies of adult patients (aged >18 years) with a first attack of AP and with a follow-up period of at least 1 month after hospital discharge were included. The main outcome measures studied were prediabetes, DM and DM treated with insulin.
Studies were excluded if they did not state that patients had no previous history of DM or prediabetes, or they did not provide data specifically on patients with no previous history of DM or prediabetes. Studies were also excluded for the following reasons: duplicate publication, retrospective study design, case report, did not use the standard definition of pancreatitis (the presence of two of the following three features: abdominal pain characteristic of AP, serum amylase and/or lipase three times the upper limit of normal and/or characteristic findings of AP on computed tomography scan), study population limited to patients who underwent pancreatic surgery such as necrosectomy or pancreatic resection, study population limited to patients with chronic, autoimmune, or hereditary pancreatitis, gestational diabetes.
Records extracted by the initial search were screened, and potentially relevant papers were retrieved and examined in more detail. Eligibility assessment was performed independently by two authors (SLMD and PPS) and any disagreement over study inclusion or exclusion was resolved by discussion with the senior author (MSP).
Data Extraction
Data were independently extracted by two authors (SLMD and PPS) using a standardised data collection form. For each eligible study the following data were extracted: (1) year conducted, (2) country/countries, (3) number of participating centres, (4) study design (eg, cohort, case-control study), (5) source of control (if applicable), (6) method of data collection, (7) criteria used to diagnose AP and classify its severity, (8) number of individuals with complete data, (9) age, (10) proportion of male individuals, (11) Body Mass Index (BMI) of these individuals, (12) aetiology of AP (ie, biliary, alcohol, other), (13) duration of follow-up period, (14) glucose homeostasis test employed, (15) criteria used to define prediabetes and DM, (16) occurrence of prediabetes, DM and DM treated with insulin, (17) number of patients with pancreatic necrosis, and (18) the presence and influence on glucose homeostasis of further attacks of AP during follow-up period. If any of these parameters were not available from the published study, the corresponding authors were contacted to provide further information.
Quality Assessment
The Newcastle-Ottawa Scale (NOS), a widely accepted tool to assess the quality of observational studies, was used. It allocates a maximum of nine points based on three aspects of study design: patient selection, comparability of study groups, and exposure and outcome of study participants. Studies were considered to be of high quality if scored 5 or greater points, and of low quality if points scored were 4 and below.
Definitions
Prediabetes was defined by specific fasting blood glucose (FBG) and/or 2 h oral glucose tolerance test (OGTT) criteria as reported by the study authors (FBG ≥5.6 mmol/L (100 mg/dL) and <7.0 mmol/L (126 mg/dL), or ≥ 6.1 mmol/L (110 mg/dL) and <7.0 mmol/L (126 mg/dL) and 2 h OGTT ≥7.8 mmol/L (140 mg/dL) and <11.1 (200 mg/dL)). DM was defined as reported by the authors of primary studies (FBG ≥7.8 mmol/L (140 mg/dL) or 2 h OGTT ≥11.1 mmol/L (200 mg/dL) or FBG ≥7.0 mmol/L (126 mg/dL)) and/or treatment with insulin, oral hypoglycaemia agents or specific dietary management. Treatment with insulin was defined as DM requiring permanent insulin therapy at follow-up (provided the mean follow-up time was equal to or greater than 3 months). Prevalence of prediabetes, DM and DM treated with insulin after AP was defined as the proportion of patients with each outcome listed after AP which was previously undiagnosed. Ideally we would have liked to report the cumulative incidence of prediabetes, DM and DM treated with insulin after AP, but given that regular universal screening for diabetes is uncommon and variable, we could not exclude pre-existing DM among those without history of diagnosed DM or prediabetes. We therefore chose to present data as prevalence of newly diagnosed DM over time since first attack of AP.
The 'time course' is defined at the prevalence of each outcome listed within certain time periods. This arbitrary includes 'up to 12 months', '12–36 month', '36–60 months' and 'more than 60 months'.
The definition of DM has changed over the years, and this is reflected in the different definitions used in the studies included in this review. The first widely accepted classification of DM was published by National Diabetes Data Group (fasting ≥7.8 h or 2 h ≥11.1 mmol/L) in 1979, WHO (fasting ≥8.0 h or 2 h ≥11.0 mmol/L) in 1980 and, in modified form, in 1985 (fasting ≥7.8 h or 2 h ≥11.1), and the next major revision occurred in 1997 by American Diabetes Association (ADA), in 1999 by WHO, and again in 2003 by ADA. The major change in the 1999 classification was the lowering of the diagnostic threshold of the fasting plasma glucose concentration. According to the 1999 WHO classification, DM was defined as a FBG of ≥7.0 mmol L (126 mg dL) or 2 h OGTT 2 h ≥11.1 mmol/L (200 mg/dL).
Studies were considered prospective if follow-up assessment was done prospectively. A binary classification was used to determine the severity of AP since the recent international multidisciplinary classification of AP severity, published by the Pancreatitis Across Nations Clinical Research and Education Alliance (PANCREA), was not used in the primary studies. This means that the study population was labelled as 'severe' if the studies reported on haemorrhagic AP, necrotising AP, or severe AP based on the Atlanta criteria. The remaining study populations were labelled as 'mild'.
Data Synthesis and Statistical Analysis
All the eligible studies were combined to give a pooled prevalence for each of the outcomes studied. If the papers reported groups of individuals with different mean follow-up times, the group with the longest follow-up was analysed for pooled prevalence. Data were pooled using a random-effects model to give the most conservative estimate. Stats Direct V.2.7.9 (StatsDirect, 2008) was used to generate forest plots of pooled prevalence with 95% CIs. Statistical heterogeneity between the studies was assessed using the I metric with cut-offs of 25, 50 and 75% to define low, moderate and high heterogeneity, respectively. Publication bias was assessed by applying the Horbold–Egger's test to funnel plots. Subgroup analysis was conducted according to mean follow-up time in the study, and they were grouped as follows: up to 12 months, between 12 and 36 months, between 36 and 60 months, more than 60 months. Sensitivity analyses constrained to studies that employed the 1999 WHO definition of DM, and to studies with severe AP, were conducted. Metaregression was used to investigate whether sex, age, duration of follow-up and aetiology of AP (proportion of biliary or alcohol AP) had a significant influence on the summary estimates. The R 2.15.2 framework with package metafor was used for metaregression analysis. Reporting of this study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines.