ABSTRACT
Background
Pancreatic exocrine insufficiency (PEI) is a maldigestion syndrome leading to symptoms and nutritional deficiencies. Prospective real‐world data on pancreatic enzyme replacement therapy (PERT) dose requirements across aetiologies are limited.
Objective
To define PERT dose requirements and treatment outcomes across PEI aetiologies in specialist European practice.
Methods
EURO‐PEI was a multicentre observational registry with predominantly prospective data collection conducted at eight pancreatic centres in five European countries between June 2021 and September 2025. Adults with clinician‐established PEI in whom PERT was initiated or actively optimised were included. The primary outcome was the optimised PERT dose associated with the achievement of a clinician‐assessed therapeutic goal: improvement or resolution of PEI‐related symptoms with normalisation or clinically relevant improvement of nutritional abnormalities.
Results
Among 1015 patients, PEI was secondary to chronic pancreatitis in 496 (48.9%), pancreatectomy in 185 (18.2%), pancreatic cancer in 174 (17.1%), and acute pancreatitis in 95 (9.4%). Median total daily optimised doses were highest after pancreatectomy (240,000 PhU, IQR 175,000–375,000) and in pancreatic cancer (210,000 PhU, 175,000–325,000), compared with chronic pancreatitis (175,000 PhU, 120,000–210,000) and acute pancreatitis (100,000 PhU, 75,000–200,000). After optimisation, the therapeutic goal was achieved in 951 patients (93.7%); of these, 446 (46.9%) required enzyme doses exceeding commonly recommended starting doses. Related adverse events occurred in 38 patients (3.7%).
Conclusions
PERT dose requirements vary substantially by PEI aetiology, supporting aetiology‐aware starting doses, early reassessment, and individualised optimisation in high‐risk groups.
Keywords: chronic pancreatitis, dose optimisation, pancreatectomy, pancreatic cancer, pancreatic enzyme replacement therapy, pancreatic exocrine insufficiency, real‐world evidence
Key Summary
-
Summarise the established knowledge on this subject
-
◦
Pancreatic exocrine insufficiency causes gastrointestinal symptoms and nutritional deficiencies across pancreatic and selected extrapancreatic diseases.
-
◦
Pancreatic enzyme replacement therapy is indicated, but real‐world dosing remains heterogeneous and aetiology‐specific dose requirements are poorly defined.
-
◦
Diagnosis and treatment monitoring require integration of symptoms, nutritional status, and pancreatic function testing in the appropriate clinical context.
-
◦
-
What are the significant and/or new findings of this study?
-
◦
In 1015 patients from eight European expert centres, optimised pancreatic enzyme doses differed markedly by aetiology.
-
◦
The highest dose requirements were observed after pancreatectomy and in pancreatic cancer, independent of patient‐level factors and centre‐level clustering.
-
◦
Nearly half of patients who achieved the therapeutic goal required doses above commonly recommended starting doses.
-
◦
Pancreatic enzyme replacement therapy was well tolerated, including at higher doses used for outcome‐driven optimisation.
-
◦
1. Introduction
Pancreatic exocrine insufficiency (PEI) is defined in recent European guidelines as reduced pancreatic exocrine secretion and/or intraluminal enzyme activity below the level required for normal digestion, leading to malabsorption, gastrointestinal symptoms, and/or nutritional deficiencies [1]. Although pancreatic diseases and pancreatectomy are the main causes, several extrapancreatic conditions may also impair effective intraluminal enzyme activity and compromise digestion [1].
PEI is therefore best regarded as a maldigestion syndrome rather than a simple pancreatic secretion defect. Its clinical expression depends not only on pancreatic enzyme output but also on gastrointestinal anatomy, intraluminal pH, enzyme–chyme mixing, compensatory non‐pancreatic digestive mechanisms, dietary intake, and individual nutritional requirements [1]. Across aetiologies, PEI may cause steatorrhoea, diarrhoea, bloating, abdominal cramps, flatulence, weight loss, and protein, fat‐soluble vitamin, and micronutrient deficiencies, with adverse effects on quality of life and longer‐term outcomes [2, 3, 4, 5, 6, 7].
Pancreatic enzyme replacement therapy (PERT), together with dietary advice and nutritional support, is the cornerstone of PEI management [1, 8, 9]. Current guidelines recommend initial adult doses of 40,000–50,000 lipase units with main meals and 20,000–25,000 with snacks, followed by individualised escalation according to response. Adherence, timing of enzyme intake, dose escalation, and acid suppression should be considered in partial responders [1, 8]. However, real‐world diagnostic and therapeutic practices remain heterogeneous, and the enzyme doses required to achieve meaningful symptomatic and nutritional improvement across PEI aetiologies are not well defined [1, 10].
EURO‐PEI was established to characterise PEI across aetiologies in specialist European practice and to define aetiology‐specific enzyme dose requirements associated with symptom control and nutritional improvement.
2. Materials and Methods
2.1. Study Design and Participants
EURO‐PEI was a multicentre observational registry with predominantly prospective data collection conducted at eight specialist pancreatic centres in Spain, Hungary, France, Italy, and the United Kingdom. All participating sites were tertiary centres with dedicated pancreatic units within gastroenterology or hepatopancreatobiliary departments and established expertise in the diagnosis and management of PEI. The registry was designed to capture routine specialist practice without protocol‐mandated diagnostic testing, PERT dosing, proton pump inhibitor (PPI) use, or follow‐up intervals. The registry was endorsed by the European Pancreatic Club, approved by the Research Ethics Committee of the University Hospital of Santiago de Compostela (2021/273), and registered in ClinicalTrials.gov (NCT05450627). Local ethics approval was obtained according to national requirements. Written informed consent was obtained from participants enrolled prospectively. In one centre, patients were identified from a pre‐existing clinical database and entered retrospectively into the registry; additional local audit approval was obtained for these data. The inclusion of these patients was planned from the outset of the registry because they fulfilled the eligibility criteria and the available clinical information was considered sufficient to complete the registry dataset.
Adults aged 18 years or older were eligible if they had a clinician‐established diagnosis of PEI and PERT was initiated or actively managed during specialist follow‐up. Patients were included between June 2021 and September 2025. Patients already receiving PERT at registry inclusion, including those entered from the pre‐existing database, were eligible if the diagnosis of PEI and the indication for ongoing treatment were confirmed by the treating specialist. In these cases, treatment outcomes and subsequent PERT optimisation were assessed from the corresponding baseline data in the source database. PEI was secondary to acute pancreatitis, chronic pancreatitis, cystic fibrosis, pancreatic cancer, pancreatectomy, or other pancreatic or extrapancreatic conditions leading to PEI.
2.2. Diagnosis, Data Collection, and Outcomes
PEI was diagnosed by treating specialists according to routine clinical practice and contemporary guideline principles, using an integrated assessment of clinical context, gastrointestinal symptoms, nutritional status, and pancreatic function testing when deemed necessary to establish or support the diagnosis. The registry did not systematically record the specific diagnostic basis for each patient. Functional testing included faecal elastase‐1 or other locally available tests. Data were entered locally into a dedicated REDCap electronic case report form by trained data managers and centrally reviewed by the global principal investigator and study coordinator. All patients were anonymised using coded identifiers to ensure data confidentiality and compliance with the European General Data Protection Regulation.
The primary outcome was the pancreatic enzyme dose, expressed in PhU (European Pharmacopoeia lipase units), required to achieve a clinician‐assessed therapeutic goal in real‐world practice: disappearance or clinically meaningful reduction of PEI‐related symptoms together with normalisation or substantial improvement of abnormal nutritional parameters. Complete resolution of all symptoms or full normalisation of all markers was not required when persistent abnormalities were judged unlikely to reflect ongoing PEI or to improve with further enzyme escalation. This pragmatic endpoint was not centrally adjudicated and reflected the dose considered clinically adequate by the treating specialist after assessment of symptoms, biochemical nutritional abnormalities, adherence, timing of enzyme intake, and plausible non‐PEI explanations for persistent abnormalities. Inadequate response was defined as persistence of symptoms and/or biochemical abnormalities considered by the managing specialist to be most likely related to ongoing maldigestion despite the current PERT regimen. Treatment failure was defined as failure to achieve the clinician‐assessed therapeutic goal after optimisation considered adequate by the treating specialist. Secondary outcomes included symptoms and nutritional deficiencies at diagnosis, initial PERT doses, PPI use during optimisation, proportion of patients achieving the clinician‐assessed therapeutic goal, persistent symptoms or nutritional deficiencies despite optimised therapy, treatment failure, and safety.
2.3. Assessment and Treatment Optimisation
Gastrointestinal symptoms were recorded before PERT initiation and during follow‐up. In patients already receiving PERT at registry inclusion, symptoms at diagnosis or treatment initiation were retrieved retrospectively from clinical documentation, and follow‐up symptoms were recorded prospectively. Nutritional assessment included body mass index and locally available biochemical markers used as indirect indicators of nutritional status, including lymphocytes, albumin, prealbumin, retinol‐binding protein, transferrin, fat‐soluble vitamins, magnesium, zinc, and selenium. Because some of these markers may be influenced by inflammation, malignancy, surgery, reduced intake, or acute illness, abnormalities were interpreted by treating specialists in the overall clinical context rather than as isolated nutrient deficiencies. PERT was initiated and escalated by treating specialists according to clinical response. In patients with an inadequate response, adherence and timing of enzyme intake were reviewed, and a PPI could be added when clinically indicated. Potential treatment‐related adverse events were recorded when the treating specialist considered a new or worsening clinical event to be possibly or probably related to PERT; dose reduction and discontinuation due to adverse events were also recorded.
2.4. Statistical Analysis
Continuous variables are summarised as mean (standard deviation) or median (interquartile range), and categorical variables as n (%). Fisher's exact, chi‐squared, Wilcoxon rank‐sum, and Kruskal–Wallis tests were used as appropriate. Secondary analyses were exploratory, and p values were interpreted descriptively without formal adjustment for multiple comparisons. Multivariable linear regression was used to evaluate factors associated with the final optimised main‐meal PERT dose. Covariates included age, sex, treatment centre, PEI aetiology, body mass index, baseline malabsorption symptoms, baseline nutritional abnormalities, PERT formulation, PPI use, faecal elastase‐1 category, and mode of data collection. Mode of data collection was entered as a binary variable to account for patients enrolled and followed prospectively and those entered retrospectively from a pre‐existing clinical database. Because optimised doses were right‐skewed, log‐transformed models were used to assess robustness.
Mixed‐effects linear regression with hospital as a random intercept assessed clustering and prescribing variability across centres; the intraclass correlation coefficient quantified residual centre‐level variability. Logistic regression models for failure to achieve the therapeutic goal used Firth's penalised likelihood correction. Complete‐case and multiple‐imputation sensitivity analyses were performed for the primary adjusted model; multiple imputation by chained equations used 20 datasets under the missing‐at‐random assumption. All statistical analyses were conducted using Stata/IC version 16.1 (StataCorp LLC, College Station, TX, USA).
3. Results
3.1. Study Population
The registry included 1015 patients with PEI. Most patients (893 [88.0%]) were enrolled at the time of PERT initiation and underwent prospective dose optimisation, while 122 (12.0%) were already receiving PERT at inclusion and were followed for outcome assessment and treatment optimisation as needed. Mean age was 61.1 years (SD 13.9), and 661 patients were male (65.1%). Enrolment was distributed across Spain (504 [49.7%]), Hungary (153 [15.1%]), France (128 [12.6%]), the UK (122 [12.0%]), and Italy (108 [10.6%]). Baseline demographic and clinical characteristics are presented in Table S1. Consecutive patients with clinician‐established PEI were enrolled, but screening failures, patients declining participation, and individuals assessed but not included because PEI was not confirmed were not prospectively recorded, precluding construction of a formal patient flow diagram.
PEI was most commonly secondary to chronic pancreatitis (496 [48.9%]), followed by pancreatectomy (185 [18.2%]), pancreatic cancer (174 [17.1%]), acute pancreatitis (95 [9.4%]), and other aetiologies (65 [6.4%]), including cystic fibrosis (n = 20), cystic neoplasms (n = 11), gastrectomy (n = 8), neuroendocrine tumours (n = 7), fatty pancreas (n = 7), and less frequent conditions (n = 12). The pancreatectomy group comprised 143 patients who had undergone pancreaticoduodenectomy, 21 total pancreatectomy, and 21 distal pancreatectomy.
3.2. Symptoms and Nutritional Deficiencies
At PEI diagnosis, 70.2% of patients reported at least one PEI‐related abdominal or intestinal symptom, with variation by aetiology (77.5% in pancreatic cancer, 75.8% in acute pancreatitis, 72.2% in chronic pancreatitis, 54.9% after pancreatectomy, and 70.8% in other aetiologies; p < 0.001; Table 1). Nutritional marker assessment was non‐systematic and subject to indication bias. Serum nutritional markers were assessed in 792 patients (78.0%); at least one biochemical nutritional abnormality was present in 49.7%, with the highest frequencies in pancreatic cancer (60.2%) and acute pancreatitis (56.6%; p = 0.005). Among patients in whom biochemical markers were measured, vitamin D deficiency was the most prevalent abnormality (75.5%), followed by magnesium deficiency (56.9%), reduced pre‐albumin (53.2%), and low transferrin (50.0%; Table 2).
TABLE 1.
Frequency of individual pancreatic exocrine insufficiency (PEI)‐related symptoms in patients included in the EURO‐PEI Registry at the time of PEI diagnosis.
| Pancreatic cancer (n = 174) | Pancreatectomy (n = 185) | Chronic pancreatitis (n = 496) | Acute pancreatitis (n = 95) | Other aetiologies (n = 65) | Total (n = 1015) | |
|---|---|---|---|---|---|---|
| Diarrhoea | 50 (28.7%) | 44 (23.8%) | 159 (32.1%) | 34 (35.8%) | 24 (36.9%) | 311 (30.6%) |
| Weight loss | 103 (59.2%) | 57 (30.8%) | 161 (32.5%) | 31 (32.6%) | 21 (32.3%) | 373 (36.8%) |
| Abdominal distention | 74 (42.5%) | 38 (20.5%) | 169 (34.1%) | 51 (53.7%) | 15 (23.1%) | 347 (34.2%) |
| Abdominal cramps | 43 (24.7%) | 9 (4.9%) | 72 (14.5%) | 34 (35.8%) | 9 (13.9%) | 167 (16.5%) |
| Bloating | 52 (29.9%) | 20 (10.8%) | 100 (20.2%) | 34 (35.8%) | 15 (23.1%) | 221 (21.8%) |
| Flatulence | 57 (32.8%) | 27 (14.6%) | 117 (23.6%) | 31 (32.6%) | 17 (26.2%) | 249 (24.5%) |
| Other symptoms | 11 (6.4%) | 15 (8.1%) | 35 (7.1%) | 7 (7.4%) | 3 (4.6%) | 71 (7.0%) |
Note: Data are shown as n (%). Percentages are calculated using the total number of patients within each aetiology as the denominator. ‘Other aetiologies’ include cystic fibrosis, cystic neoplasms, gastrectomy, neuroendocrine tumours, fatty pancreas, and other less frequent conditions. Patients may have reported more than one symptom.
TABLE 2.
Frequency of individual blood nutritional deficiencies in patients included in the EURO‐PEI Registry at the time of pancreatic exocrine insufficiency diagnosis.
| Pancreatic cancer | Pancreatectomy | Chronic pancreatitis | Acute pancreatitis | Other aetiologies | Total | |
|---|---|---|---|---|---|---|
| Lymphocyte | 16/90 (17.8%) | 20/58 (34.5%) | 23/228 (10.1%) | 12/65 (18.5%) | 1/13 (7.7%) | 72/454 (15.9%) |
| Albumin | 46/90 (51.1%) | 39/58 (67.2%) | 54/222 (24.3%) | 26/61 (42.6%) | 4/13 (30.8%) | 169/444 (38.1%) |
| Prealbumin | 39/54 (72.2%) | 27/39 (69.2%) | 54/142 (38.0%) | 18/25 (72.0%) | 4/7 (57.1%) | 142/267 (53.2%) |
| Retinol‐binding‐protein | 8/34 (23.5%) | 9/24 (37.5%) | 39/111 (35.1%) | 11/19 (57.9%) | 3/5 (60.0%) | 70/193 (36.3%) |
| Transferrin | 27/43 (62.8%) | 19/31 (61.3%) | 48/125 (38.4%) | 18/26 (69.2%) | 2/3 (66.7%) | 114/228 (50.0%) |
| Magnesium | 37/78 (47.4%) | 30/44 (68.2%) | 107/185 (57.8%) | 34/55 (61.8%) | 2/7 (28.6%) | 210/369 (56.9%) |
| Zinc | 4/36 (11.1%) | 3/15 (20.0%) | 7/101 (6.9%) | 1/14 (7.1%) | 0/3 (0%) | 15/169 (8.9%) |
| Vitamin D | 35/47 (74.5%) | 14/17 (82.4%) | 83/108 (76.9%) | 8/14 (57.1%) | 5/6 (83.3%) | 145/192 (75.5%) |
| Vitamin E | 4/33 (12.1%) | 3/12 (25.0%) | 11/103 (10.7%) | 1/9 (11.1%) | 1/4 (25.0%) | 20/161 (12.4%) |
Note: Data are presented as number of patients with deficiency of each nutritional marker related to the total number of patients in whom each individual marker was measured. Percentages are calculated using the total number of patients within each aetiology in whom the individual nutritional parameters were measured as the denominator. ‘Other aetiologies’ include cystic fibrosis, cystic neoplasms, gastrectomy, neuroendocrine tumours, fatty pancreas, and other less frequent conditions. Individual markers are reported only when measured in a sufficient number of patients for meaningful aetiology‐specific reporting; vitamin A, vitamin K, selenium, and other micronutrients were not routinely measured across participating centres and are therefore not shown.
3.3. Initial and Optimised Enzyme Dose and PPI Therapy
All patients received PERT as enteric‐coated minimicrospheres, microspheres, or minitablets. The median initial main‐meal dose was 50,000 PhU for chronic pancreatitis, pancreatic cancer, pancreatectomy, and other aetiologies, and 25,000 PhU for acute pancreatitis. Initial main‐meal and total daily doses were highest in pancreatic cancer (Figure 1). Initial main‐meal dose in pancreatic cancer was higher than after pancreatectomy (p = 0.020), chronic pancreatitis and acute pancreatitis (both p < 0.001), and other aetiologies (p = 0.019). Initial total daily dose in pancreatic cancer was also higher than after pancreatectomy, chronic pancreatitis, acute pancreatitis (all p < 0.001), and other aetiologies (p = 0.008).
FIGURE 1.
Initial and optimised pancreatic enzyme replacement therapy doses according to pancreatic exocrine insufficiency aetiology. (A) Main‐meal dose. (B) Total daily dose. Optimised dose was defined as the dose at which the clinician‐assessed therapeutic goal was achieved, based on symptom relief and normalisation or substantial improvement of nutritional status. Boxes show medians and interquartile ranges; whiskers extend to the most extreme values within 1.5 times the interquartile range, and individual points represent outliers. p values refer to within‐aetiology comparisons between initial and optimised doses. PhU, Pharmacological Units; PERT, Pancreatic enzyme replacement therapy.
Compared with the initial prescription, optimised main‐meal dose increased across all aetiologies, whereas optimised total daily dose increased in all groups except acute pancreatitis (Figure 1).
Optimised PERT requirements differed markedly across aetiological groups (Figure 1). Median optimised main‐meal dose was highest after pancreatectomy (75,000 PhU, IQR 60,000–105,000), followed by pancreatic cancer (70,000 PhU, 50,000–100,000), other aetiologies (55,000 PhU, 50,000–75,000), chronic pancreatitis (50,000 PhU, 40,000–70,000), and acute pancreatitis (30,000 PhU, 25,000–50,000). Main‐meal dose requirements after pancreatectomy were higher than in pancreatic cancer (p = 0.005), chronic pancreatitis, acute pancreatitis, and other aetiologies (all p < 0.001). Median total daily optimised dose was 240,000 PhU (IQR 175,000–375,000) after pancreatectomy, 210,000 PhU (175,000–325,000) in pancreatic cancer, 200,000 PhU (150,000–300 000) in other aetiologies, 175,000 PhU (120,000–210,000) in chronic pancreatitis, and 100,000 PhU (75,000–200,000) in acute pancreatitis. Total daily requirements were similar after pancreatectomy and in pancreatic cancer (p = 0.055), but were higher than in chronic pancreatitis, acute pancreatitis (both p < 0.001), and other aetiologies (p = 0.002; Figure 1). Among patients with PEI after pancreatectomy, median total daily optimised dose varied according to the type of surgery, being highest after total pancreatectomy (420,000 PhU, IQR 277,500–490,000), followed by pancreaticoduodenectomy (280,000 PhU, 175,000–385,000) and distal pancreatectomy (200,000 PhU, 150,000–240,000; p < 0.001).
Overall, 134 patients (13.2%) achieved the therapeutic goal after PPI introduction or escalation in addition to PERT, most frequently after pancreatectomy (21.6%; Table 3).
TABLE 3.
Evolution of proton pump inhibitor (PPI) therapy during pancreatic enzyme replacement therapy optimisation according to pancreatic exocrine insufficiency aetiology. Data are shown as n (%). Percentages are calculated using the total number of patients within each aetiology as the denominator.
| At registry entry | At time of therapeutic goal | |||||
|---|---|---|---|---|---|---|
| No PPI | Once‐daily PPI | Twice‐daily PPI | No PPI | Once‐daily PPI | Twice‐daily PPI | |
| Chronic pancreatitis | 310 (62.5%) | 154 (31.0%) | 32 (6.5%) | 282 (56.9%) | 156 (31.4%) | 58 (11.7%) |
| Acute pancreatitis | 55 (57.9%) | 33 (34.7%) | 7 (7.4%) | 53 (55.8%) | 31 (32.6%) | 11 (11.6%) |
| Pancreatic cancer | 61 (35.1%) | 82 (47.1%) | 31 (17.8%) | 73 (41.9%) | 60 (34.5%) | 41 (23.6%) |
| Pancreatectomy | 60 (32.4%) | 102 (55.1%) | 23 (12.5%) | 56 (30.3%) | 70 (37.8%) | 59 (31.9%) |
| Other aetiologies | 38 (58.5%) | 25 (38.5%) | 2 (3.1%) | 29 (44.6%) | 27 (41.5%) | 9 (13.8%) |
| Total | 524 (51.6%) | 396 (39.0%) | 95 (9.4%) | 493 (48.6%) | 344 (33.9%) | 178 (17.5%) |
Note: Other aetiologies include cystic fibrosis, cystic neoplasms, gastrectomy, neuroendocrine tumours, fatty pancreas, and other less frequent conditions.
3.4. Outcomes
The therapeutic goal was achieved with the initial enzyme dose in 570 patients (56.2%), with the highest proportion in acute pancreatitis and the lowest after pancreatectomy. After dose optimisation, 951 patients (93.7%) achieved the therapeutic goal, including 100% with acute pancreatitis, 97.0% with chronic pancreatitis, 89.8% after pancreatectomy, and 83.4% with pancreatic cancer (Table 4). Among patients who achieved the therapeutic goal, 446 (46.9%) required doses exceeding the commonly recommended starting doses. The median time from PERT initiation to achievement of the clinician‐assessed therapeutic goal was 26.5 months (IQR 10.6–72.0).
TABLE 4.
Proportion of patients achieving the therapeutic goal according to enzyme dosing strategy and PEI aetiology. Data are shown as n (%). Percentages are calculated using the total number of patients within each aetiology as the denominator.
| Initial enzyme dose (n = 1015) | Optimised enzyme dose (n = 1015) | |
|---|---|---|
| Chronic pancreatitis | 310 (62.5%) | 481 (97.0%) |
| Acute pancreatitis | 72 (75.8%) | 95 (100%) |
| Pancreatic cancer | 88 (50.6%) | 145 (83.4%) |
| Pancreatectomy | 59 (31.9%) | 166 (89.8%) |
| Other aetiologies | 41 (63.1%) | 64 (98.5%) |
| Total | 570 (56.2%) | 951 (93.7%) |
Note: Other aetiologies include cystic fibrosis, cystic neoplasms, gastrectomy, neuroendocrine tumours, fatty pancreas, and other less frequent conditions.
Despite therapy considered optimal by the managing specialist, 168 patients (17.7%) had persistent gastrointestinal symptoms and 141 (14.8%) had persistently low biochemical markers used to monitor nutritional status. Treatment failure was assigned only when persistent symptoms and/or biochemical abnormalities were considered by the managing specialist to be most likely related to insufficient response to PERT rather than to alternative non‐PEI explanations. Persistent symptoms and low biochemical markers were most frequent in patients with pancreatic cancer (33 patients [23.6%] and 34 patients [30.1%], respectively). The most frequently reported persistent symptoms were abdominal distension (7.1% of patients), diarrhoea (5.3%), and flatulence (4.8%). The most commonly affected markers were vitamin D (69.0%), pre‐albumin (61.8%), magnesium (50.4%), RBP (42.3%), and transferrin (40.4%). Together with a significant variability across centres and baseline symptoms and nutritional abnormalities, pancreatic cancer was independently associated with higher odds of treatment failure (Figure S1).
3.5. Multivariable and Sensitivity Analyses
In multivariable linear regression, PEI aetiology and treatment centre were the main variables independently associated with the final optimised main‐meal PERT dose. Compared with chronic pancreatitis, higher final doses were associated with pancreatic cancer and pancreatectomy (p < 0.001; Figure 2). Twice‐daily PPI use was also associated with higher final dose requirements (p < 0.001). Age, sex, body mass index, baseline symptoms, baseline nutritional abnormalities, PERT formulation, once‐daily PPI use, faecal elastase‐1 category, and mode of data collection were not independently associated with final dose (Figure 2).
FIGURE 2.
Factors independently associated with the final optimised main‐meal pancreatic enzyme replacement therapy dose. Multivariable linear regression analysis of variables associated with final optimised main‐meal dose in patients with pancreatic exocrine insufficiency. Sex, minimicrosphere formulation, baseline malabsorption symptoms, nutritional abnormalities, and mode of data collection were entered as binary variables. PEI aetiology, treatment centre, proton pump inhibitor use, and faecal elastase‐1 category were entered as categorical variables, whereas age and body mass index were entered as continuous variables. CI, Confidence interval; FE‐1, Fecal elastase‐1; PhU, Pharmacological Units; PPI, Proton pump inhibitor.
Centre‐level prescribing variation persisted after adjustment. In the mixed‐effects model, hospital‐level clustering accounted for 8% of residual variability in the final optimised main‐meal dose (ICC 0.08). Pancreatic cancer, pancreatectomy, and twice‐daily PPI use remained independently associated with higher dose requirements (Figure S2). Findings were consistent in the log‐transformed, complete‐case, and multiple‐imputation analyses (Tables S2 and S3), with missing data summarised in Table S4. Minimicrosphere formulation was associated with higher final doses in complete‐case and imputed analyses, but not in the primary adjusted, log‐transformed, or mixed‐effects models.
3.6. Safety
At registry closure, median PERT exposure was 26.5 months (IQR 10.6–72.0). Related adverse events occurred in 38 patients (3.7%), mostly constipation; 12 patients (1.2%) reduced dose and 6 (0.6%) discontinued PERT because of adverse events (Table S5). Adverse events were not associated with enzyme dose.
4. Discussion
EURO‐PEI provides predominantly prospective real‐world data on PEI presentation, PERT optimisation, and outcomes across a broad range of aetiologies in European specialist practice. The central finding is that enzyme dose requirements varied markedly by aetiology, with the highest optimised doses after pancreatectomy and in pancreatic cancer. Several additional findings are of clinical relevance. Almost half of patients who achieved the clinician‐assessed therapeutic goal required enzyme doses above commonly recommended starting doses; optimised total daily doses often exceeded 200,000 PhU, particularly in high‐risk aetiologies; and most patients achieved the therapeutic goal after active treatment optimisation. Finally, the aetiology‐specific differences in dose requirements persisted after adjustment for patient‐level factors, centre effects, mode of data collection, model specification, and missing baseline data.
PEI was associated with substantial symptomatic and biochemical nutritional abnormalities at diagnosis. This symptomatic and nutritional burden observed in EURO‐PEI is consistent with previous disease‐specific studies, but extends these observations across a broader PEI population [2, 11, 12, 13]. Almost half of the assessed patients had at least one biochemical nutritional abnormality, and multiple abnormalities were particularly common in pancreatic cancer and after pancreatectomy. These abnormalities should be interpreted in context because several markers may reflect inflammation, malignancy, surgery, reduced intake, or acute illness in addition to PEI‐related maldigestion. These findings support the current view of PEI as a maldigestion syndrome with systemic nutritional consequences, and reinforce guideline recommendations that diagnosis and follow‐up integrate clinical context, symptoms, nutritional status, and pancreatic function testing when clinically needed [1, 8].
The main contribution of this registry is the characterisation of aetiology‐specific enzyme dose requirements. Patients with pancreatic cancer and those after pancreatectomy required higher doses to achieve clinically meaningful symptoms and nutritional improvement than those with chronic or acute pancreatitis. These differences were not explained by age, sex, body mass index, baseline symptoms, baseline nutritional abnormalities, PERT formulation, faecal elastase‐1 category, or mode of data collection. They are biologically plausible: pancreatic cancer and pancreatectomy may combine duct obstruction, loss of pancreatic parenchyma, altered transit, biliary diversion, reduced bicarbonate secretion, low intraluminal pH, and impaired enzyme–chyme mixing [14, 15, 16, 17]. In pancreatic cancer, cachexia, systemic inflammation, chemotherapy, reduced intake, and biliary or duodenal obstruction may further increase nutritional vulnerability and make complete symptom resolution less achievable [18].
These findings have direct practical implications, although the concept of adjusting PERT according to symptoms and nutritional status is already established clinical practice. The added value of EURO‐PEI is the quantification, in a large multicentre cohort, of how often escalation beyond commonly recommended starting doses is needed and how this requirement differs by aetiology. Standard starting doses should therefore be viewed as an initial therapeutic step rather than as a target dose, especially after pancreatectomy and in pancreatic cancer. Almost half of the patients required enzyme doses above commonly recommended starting doses to achieve the therapeutic goal. This finding is clinically important because continuing standard doses despite incomplete response may leave a substantial proportion of patients undertreated. In high‐risk aetiologies such as pancreatectomy and pancreatic cancer, these data support either higher initial dosing or a lower threshold for early reassessment and escalation, rather than a uniform starting‐dose strategy for all PEI aetiologies. Importantly, dose escalation should not be regarded as exceptional in PEI management: daily dose requirements above 200,000 PhU were frequent in several aetiological groups, particularly after pancreatectomy, in pancreatic cancer, and in other complex aetiologies.
A second clinically important message is that treatment optimisation was associated with a high probability of achieving the therapeutic goal. Although only 570 patients (56.2%) achieved the therapeutic goal with the initial enzyme prescription, this proportion increased to 951 patients (93.7%) after dose optimisation. This finding supports an active, outcome‐driven approach to PEI management rather than passive continuation of the initial prescription. In practice, patients with persistent symptoms or nutritional abnormalities should undergo structured reassessment of enzyme dose, timing, adherence, dietary intake, nutritional deficiencies, and acid suppression. PPI introduction or escalation formed part of successful optimisation in 13.2% of patients, but because acid suppression was not randomly assigned and was generally used in incomplete responders, these findings should be interpreted descriptively. They support guideline‐consistent consideration of acid suppression in selected patients, not an independent causal effect of PPI therapy [1, 8, 19, 20].
Inter‐centre variability was also clinically informative. Although adjustment for hospital did not remove the association between aetiology and optimal enzyme dose, it showed that expert centres differ in prescribing thresholds and optimisation practices. This is not only a limitation of registry research; it is also a relevant finding. Even in specialist European centres, PEI management remains heterogeneous, supporting the need for more standardised, outcome‐driven algorithms that define when to reassess symptoms and nutrition, increase enzyme dose, review adherence and timing, add or intensify acid suppression, or investigate alternative causes of persistent symptoms. Minimicrosphere formulation was associated with higher final doses in complete‐case and imputed analyses, but not in the primary adjusted, log‐transformed, or mixed‐effects models, suggesting that this apparent association was largely explained by centre‐specific prescribing practices.
Persistent symptoms or nutritional deficiencies after apparent optimisation should not be interpreted automatically as failure of enzyme therapy. In pancreatic cancer, after pancreatectomy, and in other complex conditions, residual morbidity may reflect altered anatomy, reduced intake, cachexia, chemotherapy toxicity, small intestinal bacterial overgrowth, bile acid malabsorption, altered motility, diabetes, micronutrient depletion, or inadequate supplementation. Accordingly, PEI management should be iterative and multidisciplinary, combining PERT optimisation with dietary assessment, supplementation, and evaluation of competing causes of symptoms or deficiencies.
PERT was well tolerated in this large real‐world cohort, including at the higher doses used during optimisation. Adverse events were infrequent, mostly mild, and rarely led to dose reduction or discontinuation. This safety profile is clinically relevant because concerns about high‐dose PERT, uncertainty about dose ceilings, and insufficient follow‐up contribute to undertreatment [21, 22, 23]. In the context of the present findings, tolerability data support clinically guided dose escalation when symptoms or nutritional abnormalities persist, provided that response, adherence, and safety are monitored.
The strengths of EURO‐PEI include its predominantly prospective data collection, large sample size, geographical diversity, inclusion of multiple aetiologies, long treatment exposure, and standardised data capture across expert centres. The main limitations are those inherent to observational registry studies. Treatment was not randomly assigned, the diagnostic criteria and nutritional assessment were not fully standardised, and some degree of misclassification and selection bias cannot be excluded. In addition, PERT dose optimisation was guided by local clinical practice, and treatment success was determined by the treating physician rather than by central adjudication or validated patient‐reported outcome measures. A placebo effect on symptom reporting cannot be excluded. In addition, the registry did not systematically capture dietary macronutrient intake, meal composition, or validated symptom questionnaires. Faecal elastase‐1 measurements were unavailable for a substantial proportion of patients, reflecting real‐world clinical practice, in which functional testing is not always considered necessary in patients with a high pre‐test probability of PEI. Residual confounding factors such as disease severity, diet, adherence, anatomy, cancer stage, chemotherapy, diabetes, supplementation, and centre‐level practice cannot be excluded. The cohort was drawn from expert centres, which supports data quality but may limit generalisability to non‐specialist settings. Therefore, aetiology should be interpreted as an important, clinically accessible determinant of dose requirements, but not as the only factor driving individual enzyme needs. These limitations preclude causal inference but do not weaken the central descriptive message: PEI management requires aetiology‐aware, outcome‐driven dose optimisation.
In conclusion, EURO‐PEI shows that enzyme requirements differ substantially by PEI aetiology in European specialist practice. Nearly half of patients who achieved the therapeutic goal required enzyme doses above commonly recommended starting doses, optimised daily requirements frequently exceeded 200,000 PhU, and most patients achieved the therapeutic goal after structured treatment optimisation. Pancreatectomy and pancreatic cancer were associated with the highest optimised dose requirements. These findings support an aetiology‐aware approach to initial dosing and early escalation, particularly in high‐risk groups, while recognising that individual dose requirements remain multifactorial and must be guided by symptoms, nutritional assessment, diet, adherence, anatomy, acid suppression, and competing causes of persistent abnormalities. These data provide real‐world evidence to inform future guideline refinement and interventional studies of structured dose‐optimisation algorithms.
Author Contributions
All authors fulfilled the ICMJE and UEG Journal criteria for authorship. All authors contributed to patient inclusion and data acquisition, critically revised the manuscript for important intellectual content, approved the final version for publication, and agreed to be accountable for all aspects of the work, including the accuracy and integrity of the data. In addition to these contributions, J.E.D.‐M. was responsible for conceptualisation, study design, supervision, data interpretation, and drafting and finalisation of the manuscript. Y.D.‐N. performed the statistical analysis and contributed to data interpretation, and X.M.‐S. designed the electronic database.
Funding
The study was supported by an unrestricted grant from Viatris. The funder had no role in the study design, data collection, analysis, or interpretation, had no access to the study data, and had no role in the writing of the manuscript. All authors had full access to the data and accept responsibility for the decision to submit for publication.
Ethics Statement
The registry was approved by the Research Ethics Committee of the University Hospital of Santiago de Compostela (2021/273). Local ethics approval was obtained according to national requirements. Written informed consent was obtained from participants enrolled prospectively.
Conflicts of Interest
J.E.D.‐M. has received honoraria from Viatris, Abbott, Pangenix and Amgen for lectures, educational programmes and participation in advisory boards. His research team has received unrestricted research grants from Viatris. Y.D.‐N. has received honoraria from Viatris and Abbott for lectures and educational activities. V.R. has received honoraria from Mayoly Spindler for lectures and participation in symposia. M.E.P. has received honoraria from Viatris, Nutricia Clinical Care, Essential Pharmaceuticals, and Normark Pharma for lectures and participation in expert panels and consultation. G.C. has served as a consultant for Amgen, Boston Scientific, Pangenix, Viatris. K.J.R. has received honoraria from Viatris and Abbott for lectures. His research team has received unrestricted research grants from Viatris. G.d.M. has received consultancy honoraria from Zenas Biopharma. J.L.‐N. received honoraria from Viatris and Abbott for educational activities. J.I.‐G. received honoraria from Viatris and Abbott for educational activities. Other authors (X.M.‐S., P.H., F.B., D.R.‐C., G.L., O.E., S.S., A.B., S.P.‐B.) have nothing to declare.
Supporting information
Supporting Information S1
Figure S1: Factors associated with failure to achieve the therapeutic goal in patients with pancreatic exocrine insufficiency.
Figure S2: Factors independently associated with final optimised main‐meal pancreatic enzyme replacement therapy dose after accounting for inter‐centre variability.
Table S1: Baseline demographic and clinical characteristics of the study population.
Table S2: Adjusted multivariable linear regression analysis using log‐transformed final main‐meal pancreatic enzyme replacement therapy dose.
Table S3: Sensitivity analysis for the primary outcome (final dose main meals) using complete‐case and multiple‐imputation multivariable linear regression analyses.
Table S4: Missing data for variables included in multivariable and mixed‐effects regression analyses.
Table S5: Type and severity of adverse events related to pancreatic enzyme replacement therapy according to pancreatic exocrine insufficiency aetiology.
Acknowledgements
The authors thank all patients and participating centres for their contribution to the EURO‐PEI Registry. We thank Rachel Thompson, Department of Nutrition and Dietetics, Royal Surrey NHS Foundation Trust, Guildford, UK for her support in the study. The authors acknowledge the use of ChatGPT (GPT‐5.5 Thinking, OpenAI) and Claude (Anthropic) solely to assist with spelling, grammar, and language clarity. All outputs were critically reviewed by the authors, who take full responsibility for the final content of the manuscript.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Dominguez‐Muñoz J. E., Vujasinovic M., De La Iglesia D., et al., “European Guidelines for the Diagnosis and Treatment of Pancreatic Exocrine Insufficiency: UEG, EPC, EDS, ESPEN, ESPGHAN, ESDO, and ESPCG Evidence‐Based Recommendations,” United European Gastroenterology Journal 13, no. 1 (2025): 125–172, 10.1002/ueg2.12674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Lindkvist B., Phillips M. E., and Domínguez‐Muñoz J. E., “Clinical, Anthropometric and Laboratory Nutritional Markers of Pancreatic Exocrine Insufficiency: Prevalence and Diagnostic Use,” Pancreatology 15, no. 6 (2015): 589–597, 10.1016/j.pan.2015.07.001. [DOI] [PubMed] [Google Scholar]
- 3. Lindkvist B., Domínguez‐Muñoz J. E., Luaces‐Regueira M., Castiñeiras‐Alvariño M., Nieto‐Garcia L., and Iglesias‐Garcia J., “Serum Nutritional Markers for Prediction of Pancreatic Exocrine Insufficiency in Chronic Pancreatitis,” Pancreatology 12, no. 4 (2012): 305–310, 10.1016/j.pan.2012.04.006. [DOI] [PubMed] [Google Scholar]
- 4. Ramai D., Facciorusso A., Maida M., et al., “Prevalence of Osteopathy in Chronic Pancreatitis: A Systematic Review and Meta‐Analysis,” Clinical and Translational Gastroenterology 14, no. 8 (2023): e00623, 10.14309/ctg.0000000000000623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Matsumoto R., Kikuta K., Takikawa T., et al., “Skeletal Muscle Mass and Function Are Affected by Pancreatic Atrophy, Pancreatic Exocrine Insufficiency and Poor Nutritional Status in Patients With Chronic Pancreatitis,” Pancreatology 24, no. 2 (2024): 197–205, 10.1016/j.pan.2024.01.002. [DOI] [PubMed] [Google Scholar]
- 6. de la Iglesia‐Garcia D., Vallejo‐Senra N., Iglesias‐Garcia J., López‐López A., Nieto L., and Domínguez‐Muñoz J. E., “Increased Risk of Mortality Associated With Pancreatic Exocrine Insufficiency in Patients With Chronic Pancreatitis,” Journal of Clinical Gastroenterology 52, no. 8 (2018): e63–e72, 10.1097/MCG.0000000000000917. [DOI] [PubMed] [Google Scholar]
- 7. de la Iglesia D., Vallejo‐Senra N., López‐López A., et al., “Pancreatic Exocrine Insufficiency and Cardiovascular Risk in Patients With Chronic Pancreatitis: A Prospective, Longitudinal Cohort Study,” Journal of Gastroenterology and Hepatology 34, no. 1 (2019): 277–283, 10.1111/jgh.14460. [DOI] [PubMed] [Google Scholar]
- 8. Whitcomb D. C., Buchner A. M., and Forsmark C. E., “AGA Clinical Practice Update on the Epidemiology, Evaluation, and Management of Exocrine Pancreatic Insufficiency: Expert Review,” Gastroenterology 165, no. 5 (2023): 1292–1301, 10.1053/j.gastro.2023.07.007. [DOI] [PubMed] [Google Scholar]
- 9. De La Iglesia‐García D., Huang W., Szatmary P., et al., “Efficacy of Pancreatic Enzyme Replacement Therapy in Chronic Pancreatitis: Systematic Review and Meta‐Analysis,” Gut 66, no. 8 (2017): 1354.1–1355, 10.1136/gutjnl-2016-312529. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. De Rijk F. E. M., Van Veldhuisen C. L., Besselink M. G., et al., “Diagnosis and Treatment of Exocrine Pancreatic Insufficiency in Chronic Pancreatitis: An International Expert Survey and Case Vignette Study,” Pancreatology 22, no. 4 (2022): 457–465, 10.1016/j.pan.2022.03.013. [DOI] [PubMed] [Google Scholar]
- 11. Min M., Patel B., Han S., et al., “Exocrine Pancreatic Insufficiency and Malnutrition in Chronic Pancreatitis: Identification, Treatment, and Consequences,” Pancreas 47, no. 8 (2018): 1015–1018, 10.1097/MPA.0000000000001137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kroon V. J., Daamen L. A., Tseng D. S. J., et al., “Pancreatic Exocrine Insufficiency Following Pancreatoduodenectomy: A Prospective Bi‐Center Study,” Pancreatology 22, no. 7 (2022): 1020–1027, 10.1016/j.pan.2022.08.002. [DOI] [PubMed] [Google Scholar]
- 13. Percy K. L., Ooi Y. X., Nahm C. B., et al., “Defining the Rate of Nutritional and Metabolic Derangements After Pancreatic Resection,” Pancreatology 22, no. 7 (2022): 1028–1034, 10.1016/j.pan.2022.07.014. [DOI] [PubMed] [Google Scholar]
- 14. Domínguez‐Muñoz J. E., de la Iglesia‐García D., Nieto‐García L., et al., “Endoscopic Pancreatic Drainage Improves Exocrine Pancreatic Function in Patients With Unresectable Pancreatic Cancer: A Double‐Blind, Prospective, Randomized, Single‐Center, Interventional Study,” Pancreas 50, no. 5 (2021): 679–684, 10.1097/MPA.0000000000001817. [DOI] [PubMed] [Google Scholar]
- 15. Powell‐Brett S., Halle‐Smith J. M., Hall L. A., Hodson J., Phillips M. E., and Roberts K. J., “Comprehensive, Long‐Term Evaluation of Pancreatic Exocrine Insufficiency After Pancreatoduodenectomy,” Pancreatology 24, no. 2 (2024): 298–305, 10.1016/j.pan.2023.11.016. [DOI] [PubMed] [Google Scholar]
- 16. Kanwat S., Singh H., Sharma A. K., et al., “Pancreatic Dysfunction and Reduction in Quality of Life Is Common After Pancreaticoduodenectomy,” Digestive Diseases and Sciences 68, no. 7 (2023): 3167–3173, 10.1007/s10620-023-07966-6. [DOI] [PubMed] [Google Scholar]
- 17. Downs E. M., Eaton A., Witkowski P., et al., “Nutritional Risks in Patients Undergoing Total Pancreatectomy Islet Autotransplantation in the POST Consortium,” Journal of Gastrointestinal Surgery 27, no. 9 (2023): 1893–1902, 10.1007/s11605-023-05770-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Falconer J. S., Fearon K. C., Plester C. E., Ross J. A., and Carter D. C., “Cytokines, the Acute‐Phase Response, and Resting Energy Expenditure in Cachectic Patients With Pancreatic Cancer,” Annals of Surgery 219, no. 4 (1994): 325–331, 10.1097/00000658-199404000-00001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Domínguez‐Muñoz J. E., Iglesias‐García J., Iglesias‐Rey M., and Vilariño‐Insua M., “Optimising the Therapy of Exocrine Pancreatic Insufficiency by the Association of a Proton Pump Inhibitor to Enteric Coated Pancreatic Extracts,” Gut 55, no. 7 (2006): 1056–1057, 10.1136/gut.2006.094912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. DiMagno E. P., Malagelada J. R., Moertel C. G., and Go V. L., “Prospective Evaluation of the Pancreatic Secretion of Immunoreactive Carcinoembryonic Antigen, Enzyme, and Bicarbonate in Patients Suspected of Having Pancreatic Cancer,” Gastroenterology 73, no. 3 (1977): 457–461, 10.1016/s0016-5085(19)32123-7. [DOI] [PubMed] [Google Scholar]
- 21. Forsmark C. E., Tang G., Xu H., Tuft M., Hughes S. J., and Yadav D., “The Use of Pancreatic Enzyme Replacement Therapy in Patients With a Diagnosis of Chronic Pancreatitis and Pancreatic Cancer in the US Is Infrequent and Inconsistent,” Alimentary Pharmacology & Therapeutics 51, no. 10 (2020): 958–967, 10.1111/apt.15698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Barkin J. A., Harb D., Kort J., and Barkin J. S., “Real‐World Patient Experience With Pancreatic Enzyme Replacement Therapy in the Treatment of Exocrine Pancreatic Insufficiency,” Pancreas 53, no. 1 (2024): e16–e21, 10.1097/MPA.0000000000002273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kadaj‐Lipka R., Monica M., Stożek‐Tutro A., Ryś P., and Rydzewska G., “Pancreatic Enzyme Replacement Therapy in Pancreatic Exocrine Insufficiency—Real‐World’s Dosing and Effectiveness: A Systematic Review,” Digestive Diseases and Sciences 70, no. 7 (2025): 2270–2284, 10.1007/s10620-025-09011-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information S1
Figure S1: Factors associated with failure to achieve the therapeutic goal in patients with pancreatic exocrine insufficiency.
Figure S2: Factors independently associated with final optimised main‐meal pancreatic enzyme replacement therapy dose after accounting for inter‐centre variability.
Table S1: Baseline demographic and clinical characteristics of the study population.
Table S2: Adjusted multivariable linear regression analysis using log‐transformed final main‐meal pancreatic enzyme replacement therapy dose.
Table S3: Sensitivity analysis for the primary outcome (final dose main meals) using complete‐case and multiple‐imputation multivariable linear regression analyses.
Table S4: Missing data for variables included in multivariable and mixed‐effects regression analyses.
Table S5: Type and severity of adverse events related to pancreatic enzyme replacement therapy according to pancreatic exocrine insufficiency aetiology.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.