Magnitude of pancreatic exocrine insuciency and associated factors among adult diabetic patients attending Madda Walabu University Goba referral hospital, south east Ethiopia, 2019

Background: Diabetes mellitus is a group of metabolic disorders which is characterized by increased blood glucose level. It causes a lot of systemic complications including pancreatic atrophy that leads to pancreatic exocrine insuciency. It is more common among type 1 diabetic than type 2 diabetic patients. The aim of this study was to assess magnitude of pancreatic exocrine insuciency and associated factors among diabetic patients attending Madda Walabu University Goba referral hospital, 2019. Materials and method: An institution based cross sectional study was conducted on 286 diabetic patients during study period. The study participants were selected by systematic random sampling technique among diabetic patients attending Madda Walabu University Goba referral hospital internal medicine department. The demographic data of the study participants were collected by face to face interview. Serum and urine samples were collected from each patient and analyzed to check pancreatic exocrine insuciency and factors associated with it. The raw data were entered in to Epi Data V 3.0.2 and exported in to SPSS V 25 for analysis. Descriptive analysis and multiple logistic regression was done and variables with 95% condence interval and p-value less than 0.05 were used to declare statistical signicance. Results: The study participants were 250 (87.4%) type 2 diabetic and 36 (12.6%) type 1 diabetic patients. Among them 21.3% suffered from pancreatic exocrine insuciency. Amylase insuciency found in 44.4% of type 1 and 16.0% in type 2 diabetics. But, lipase insuciency was seen in 41.7% of type 1 and 16.0% in type 2 diabetics. The pancreatic exocrine insuciency was higher among type 1 (44.44%) than type 2 (18.0%). Smoking habit, alcohol consumption, duration of disease, usage of hypoglycemic agents, ketosuria and type of diabetics were signicantly associated with pancreatic exocrine insuciency. Conclusion: There was an alteration of pancreatic exocrine enzyme secretion among both type 1 and type 2 diabetic patients. The habit of alcohol consumption, smoking and increase in the duration of the disease intensify the pancreatic insuciency.


Introduction
Diabetes mellitus (DM) is a collection of metabolic derangement manifested by hyperglycemia and it has typical systemic complications such as, cardiovascular disease, retinopathy, nephropathy and neuropathy (1). The pancreas has dual function that serves as both endocrine and exocrine glands. As endocrine gland, it produces insulin hormone which regulate blood glucose level and as an exocrine gland it releases digestive enzymes in to gut (2).
Pancreas' vital role in macronutrient digestion is to secrete digestive enzymes in to the duodenum. The pancreatic acinar cells committed to synthesis of proteolytic, amylolytic, lipolytic, and nuclease digestive enzymes. Neuropathic effect of diabetes mellitus halts this activity of pancreas, that resulted in pancreatic exocrine insu ciency (PEI) and additionally, lack of insulin causes acinar cell atrophy.
Maldigestion of nutrients due to pancreatic exocrine insu ciency resulted in malnutrition (3).
The prevalence of pancreatic exocrine insu ciency in DM is high, ranging from 26-44% and from 12-20% in T1DM and T2DM respectively (4). Diabetic neuropathy is one of chronic diabetic complication that resulted in pancreatic exocrine insu ciency. Pancreatic exocrine insu ciency was reported in 52.4% of diabetes mellitus. T1DM patients were more suffered from PEI than T2DM with a magnitude of 51% and 32% respectively (5). Pancreatic cell death leads to altered pancreatic digestive enzyme synthesis. Accordingly, 28% of T2DM and all of them (100%) had serum amylase activity < 23 U/L and serum lipase activity < 5 U/L which is below reference range. The decrement of amylase and lipase activity linked to altered exocrine pancreatic function (6).
Although a lot of study conducted on DM little was elaborated on diabetic pancreatic insu ciency. According to systematic review done in 2015, PEI was more common in both type1 and type2 DM with a prevalence of (25-74%) and (28-54%) respectively (7). In Germany 51% of T1DM and 35% of T2DM patients were suffered from PEI. As the duration of disease increases, the PEI also increases (8). Another clinical review conducted in 2011 on pancreas and DM relation indicated that 45-74% of T1DM and 28-36% of T2DM patients had exocrine pancreatic insu ciency (9).
The accumulation of fat in a pancreas due to metabolic syndrome (MetS) leads to pancreatic exocrine insu ciency. There was a reduction of Serum amylase activity in lean subjects as the progression of nonalcoholic fatty liver disease (NAFLD) (10). As the blood glucose level increases, the pancreatic capacity to secrete digestive enzyme decreases; resulting in maldigestion of carbohydrates, lipids and proteins. Therefore, analysis of serum amylase activity shows the progress of diabetic related complication (11).
The prevalence of PEI in prediabetes alone and diabetic alone is 41% and 39% respectively. Not only this but also, newly diagnosed DM and prediabetes are suffered from PEI with a prevalence of 40%. After acute pancreatitis, 39% of them suffered from PEI (12).
Even though su cient studies were not available in Ethiopia on pancreatic exocrine insu ciency due to DM, it was estimated that in 2035, 5.5% of Ethiopian population will suffer from DM. The increment of the diabetic patients from 4.4% in 2013 to 5.5% in 2035 needs a great attention to manage the complications of DM (13). So, the aim of this study was to assess the pancreatic exocrine insu ciency and associated factors among diabetic patients attending Madda Walabu University Goba referral hospital.

Materials And Method
A cross sectional study was conducted at Madda Walabu University Goba referral hospital (MWU GRH), Bale zone, South East Ethiopia which is 445 km far away from Addis Ababa from April 15 -May 30, 2019 on adult diabetic patients. All type 1 and type 2 diabetic patients of age 18 years and above visiting MWU Goba referral hospital were included. But diabetic patients those who were critically ill (in coma), diabetic patients with known pancreatic disease before DM onset and diabetic patients with salivary gland in ammation was excluded by card review and after physical inspection done by physician.
The sample size was calculated by using a single population proportion formula by considering the following assumptions. Since the prevalence of pancreatic exocrine insu ciency among diabetics in Ethiopia was not available P = 50% and margin of error d = 0.05 with 95% con dence interval was used.
The study unit was selected from the study population by systematic random sampling. There were 869 diabetic patients registered for attending MWU GRH for diabetic follow up and the sample size is 294. Therefore, the sampling interval was: The starting point was randomly selected among 1, 2 and 3; then those who registered every three intervals was participated in the study.
Interviewer-administered semi-structured questionnaire was adapted from different articles and guidelines those related to this study. Then the adapted tool was translated to the local languages Afan Oromo and Amharic languages and then back to English to ensure its consistency and accuracy. The questionnaire was contained socio-economic and demographic characteristic of the respondent, his/her health condition, life style and laboratory result registration format. The data collectors were trained on the objective of the study, data collection tool, approach to the interviewees, details of interviewing techniques, respect and maintaining privacy and con dentiality of the respondents before going to the data collection.
Data were collected by trained nurses by using a semi structured questionnaire through face to face interview and the type of DM and the type of medication they used were obtained by reviewing their card. The biological sample was collected and analyzed by the laboratory technologists according to standard operating procedure (SOP). Accordingly, 5ml of blood were collected from each study participants, then it was allowed to clot for 30 minute. Then, the cells were separated from serum by centrifuging the blood at 3000 rpm for 5 minute. The serum sample was analyzed by spectrophotometer using different reagents for each parameter. From each study participants 10ml of urine was collected and analyzed after centrifugation at 3000 rpm for 5 minute. The qualitative tests were done by immersing dry reagent dipstick in to urine and urine creatinine concentration and urine amylase activity was done by spectrophotometer.
Pancreatic exocrine insu ciency was the dependent variable and age, sex, residence, monthly income, occupation, and educational level were the socio economic independent variables. In addition to this blood pressure, type of DM, type of medication, age of onset, duration of the disease, smoking habit, alcohol use and BMI were included as life style and health condition independent variables. Serum analytes and calculated parameters such as FBS level, GFR, serum creatinine, serum urea and BUN were also included.
After checking the completeness of the data manually, the collected raw data were entered, cleaned and checked by Epi data version 4.0.2 and then it was exported to SPSS version 25 statistical packages for analysis. Descriptive analysis of different variables was done and variables with p-value <0.25 in binary logistic regression were used as a candidate for multiple logistic regression. Finally, the variables in multiple logistic regression that have signi cant association were identi ed by calculating odds ratio, with 95% con dence interval and p-value less than 0.05 was used to declare statistical signi cance.
To avoid the possibility of sampling bias, a minimum sample size su cient to study the objectives of the study was determined. Pretest was done on 5% of the sample size (21 individuals) at Bale Robe Hospital.
The collected data were checked for the completeness, clarity and consistency on the daily basis. Supervisor and principal investigator were closely follow both biological sample and information data collection process. After data collection, data were edited, coded, cleaned and some inconsistencies were checked in order to assess the quality of data.
Pancreatic exocrine insu ciency: is de ned as inability of pancreas to secrete the digestive enzymes that result in maldigestion of macromolecules. It is manifested by either serum amylase activity <28 U/l (normal range 28-100 U/l) or serum lipase activity <13 U/l (normal range 13-60 U/l) (14).
Five ml of venous blood was collected from each study participants in to tube with clot activator and the cells were separated from the serum within 30 minutes in order to prevent glycolysis that could be taken place by blood cells. Hemolysed sample was discarded because, it disturbs the absorbance of the solution during spectrophotometric reading. If the reading was above the measuring capacity of the spectrophotometer, the serum was diluted and the result was multiplied by dilution factor prior to reporting/registration (15).
Ethical clearance was obtained from institutional review board of Jimma University Institute of Health. Formal letter was obtained from Jimma University Institute of health to MWU GRH and the permission to conduct the study was obtained from chief executive of MWU GRH. Then both written and oral consent was obtained from each study participant. The name of the study participants was omitted from the questionnaire and the registration number/code was used to ensure the con dentiality. All abnormal laboratory result was linked to internal medicine department in the hospital for appropriate management of the study participants.
Almost more than half 158 (55.2%) of the study participants were only hypoglycemic agent users and about 63 (22.0%) were used both hypoglycemic agents and insulin. Regarding family history of diabetics 86 (30.1%) had family history of previous diabetics. Most of the study participants 195 (68.2%) didn't use antihyperlipemic agents (refer Table 1).

Factors associated to pancreatic exocrine insu ciency
Each independent variable entered in to binary logistic regression with pancreatic exocrine insu ciency as a dependent variable separately. Its result illustrated on the following table (Table 4). Regarding the duration of disease, the odds of being pancreatic exocrine insu cient among diabetic patients who had diabetic duration 11 to 15 years and 16 to 20 years were 5.59 (CI: 1.71, 18.30 P-value = 0.004) and 10.25 (CI: 2.53, 41.56 P-value = 0.001) times more likely to suffer from PEI respectively in comparison to diabetic patients below ve years' duration of disease. But the there was no signi cant difference between of disease 6 to 10 years and under ve years' duration.
Type 1 diabetics were 7.04 (AOR CI: 1.83, 27.04 P-value = 0.005) times more likely to suffer from PEI than T2DM. Additionally, the odds of patients with ketosuria were 6.85 (CI: 2.93, 15.98 P-value = 0.000) times more likely to suffer from PEI than those who didn't have ketosuria. Hypoglycemic agent users were more likely to develop PEI with odds ratio of 3.46 (CI: 1.04, 11.53 P-value = 0.043) as compared to non-users. Type 1 diabetics were 7.04 (95% CI: 1.86, 27.04) more likely to suffer from PEI than T2DM. It may be due to autoimmune destruction of beta cells which led to severe exocrine cell atrophy (8,25). Hypoglycemic agent users were 3.46 (95% CI: 1.04, 11.53) more likely to suffer from PEI than non-users. On the other hand, even though it is not statistically signi cant, Gonzalez et al, (2010) reported that metformin users and sulfonylurea users were more likely to develop acute pancreatitis which later on led to PEI. Such disparity might be due to study design difference and sample size inequality (23).
In current study, the odds of alcohol users were 3.40 (95% CI: 1.33, 8.65) times more likely to suffer from PEI than non-alcohol drinkers. This situation may happen due to generation of toxic metabolite and free radicals during alcohol metabolism. According to Nakajima K et al, (2011) report alcohol consumption negatively correlated with serum amylase level (β coe cient= -0.07) and signi cant difference between quartiles of serum amylase with alcohol consumption (26). The odds of diabetic patients living in rural area were 5.38 (95% CI: 2.42, 14.22) times more likely to develop PEI than urban settlers. It may be due to lack access for glycemic checkup which resulted in poor glycemic control.

Conclusion
According to this study there is an evidence for the alteration of pancreatic exocrine secretion of digestive enzymes. It shows that PEI was more prevalent and more severe among T1DM than T2DM. Therefore, PEI has to be considered as the one of diabetic complications. About one fth of diabetic patients suffered from pancreatic enzymes secretion derangement. The pancreatic enzymes secretion was signi cantly lowered by hypoglycemic agent usage, ketosuria, habit of cigarette smoking and alcohol drinking habit of the diabetic patient in the present study.