Trajectories of Pain Processing in Recurrent Acute and Chronic Pancreatitis: A Longitudinal Quantitative Sensory Testing Study.

This study was designed by S.S.O., A.M.D., J.B.F. and S.N. The experiments were performed by L.D., C.S.K. and I.M.L. The data were analysed by L.K., and the results were examined by all co‐authors. L.K. prepared the manuscript, supervised by S.S.O. All authors have approved the final version of the manuscript.

The authors have nothing to report.

This was a prospective observational study conducted from 2019 to 2024 at two tertiary care centres in Denmark (Aalborg University Hospital and Hvidovre Hospital). We enrolled patients with RAP or CP from the outpatient clinics, with annual follow‐up for up to 3 years. The full study protocol has been published previously (Novovic et al.  2019 ). Inclusion criteria included age between 18 and 75 years. RAP was defined as more than one episode of acute pancreatitis as per revised Atlanta criteria, with episodes separated by more than 30 days, and no CP at the time of inclusion. A CP diagnosis was based on the M‐ANNHEIM criteria for definitive CP (Schneider et al.  2007 ). For a complete overview of the study inclusion criteria, see the study protocol (Novovic et al.  2019 ). The study was conducted in accordance with the Danish national legislation on health and the Helsinki Declaration. It has been approved by the Danish Data Protection Agency (VD‐2018‐298) and the Regional Committee on Health Research Ethics (Journal‐no.: H‐18017705). Prior to inclusion, all participants provided written informed consent. Participation was voluntary, and participants could withdraw consent at any time without consequences. All data were handled confidentially. At baseline, data were collected on gender, age, current alcohol use and smoking habits and pancreatitis‐specific factors, including disease duration, M‐ANNHEIM risk factors (Schneider et al.  2007 ), history of acute pancreatitis, presence of pain, information on analgesic treatment and presence of exocrine and endocrine insufficiency. Exocrine pancreatic insufficiency was defined as a faecal elastase level below 200 μg/g combined with evidence of malabsorption, such as low vitamin D levels or weight loss. A diagnosis of diabetes was based on the American Diabetes Association criteria ( 2025 ). Smoking was characterised as never, active, or previous, depending on current status. Alcohol misuse was defined as an alcohol consumption of more than 10 units of alcohol per week. All risk factors were recorded retrospectively based on clinical history prior to diagnosis. At baseline, psychological factors were assessed using validated self‐report instruments. The Hospital Anxiety and Depression Scale (HADS) (Bjelland et al.  2002 ) consists of 14 items divided into anxiety and depression subscales, each scored independently, and is widely used in medical populations to detect clinically relevant affective symptoms without being confused with somatic disease burden. Scores ≥ 8 were used as a cutoff for both the depression and the anxiety score, similar to what has been used in other gastrointestinal diseases (Snijkers et al.  2021 ). Pain‐related cognitions were evaluated using the Pain Catastrophizing Scale (PCS) (Osman et al.  1997 ), a 13‐item questionnaire measuring rumination, magnification and helplessness, with higher scores indicating greater catastrophic thinking about pain. A cutoff for pain catastrophising was defined as scores ≥ 30, as recommended by the user manual (Sullivan  1995 ). At baseline and at all subsequent annual follow‐up visits, patients completed PROMs capturing pain intensity, pain‐related interference and health‐related quality of life. General quality of life was assessed using the EORTC‐QLQ‐C30 (Fayers and Bottomley  2002 ), a questionnaire comprising multi‐item functional and symptom scales and a global health status score. Pain intensity and pain interference were evaluated using the modified Brief Pain Inventory short form (mBPI‐SF) (Mendoza et al.  2006 ), which quantifies current, worst and average pain, as well as the degree to which pain affects daily activities and functioning. A severity score is calculated as an average of four pain intensity scores (worst, least, average and current). An interference score is calculated as an average of seven subscores (general activity, mood, walking, work, relations, sleep and enjoyment of life). Together, these PROMs enabled multidimensional evaluation of patients' psychological state, pain experience and quality of life throughout the study period. The study design is shown in detail in Figure  S1 . Patients underwent the P‐QST assessments at baseline and follow‐up visits. Details on the P‐QST protocol have been published previously (Phillips et al.  2020 ); a brief description of the different P‐QST procedures and parameters is provided below Pressure stimulation: Pressure pain detection thresholds (pPDT) were assessed at C5 (clavicle), T10 dorsal (pancreatic viscerotome), T10 ventral (upper abdomen pancreatic viscerotome), L1 (iliac crest) and L4 (quadriceps) on the dominant side using an electronic pressure algometer with a steadily increasing pressure rate of 30 kPa/s (Algometer Type II, SOMEDIC Electronic, Solna, Sweden). pPDTs were summed to generate an overall pain sensitivity score. A pPDT index, representing segmental pressure hyperalgesia, was calculated by dividing the sum of pPDTs at pancreatic dermatomes (T10 ventral and dorsal) by the sum at reference sites (C5, L1, L4). Cold pressor test: The dominant hand was immersed in ice‐chilled water at 2°C (approximately 36°F) for a maximum of 120 s, with endurance time recorded (seconds). Pain intensity was rated every 10 s on a 0–10 numeric rating scale (NRS). Conditioned pain modulation (CPM): CPM, considered an indirect measure of descending pain modulation, reflecting inhibitory signals from the brainstem that dampen nociceptive input at the spinal cord, was induced using the cold pressor test as a conditioning stimulus and assessed by measuring pPTT at L4 before and after conditioning. The relative change (%) in pPTT was calculated as a measure of CPM efficiency. Temporal summation: Temporal summation, considered a proxy of neuronal excitability, was assessed using an 8 mN pin‐prick device (PinPrick Stimulatoren, MRC Systems GmbH, Heidelberg, Germany). Pain intensity was rated on a 0–10 NRS after a single stimulation as well as after 10 repetitive (1 Hz) stimulations on the ventral pancreatic dermatome and the dominant forearm. The difference in pain intensity scores between the first and tenth stimulus was used as a measure for temporal summation. Pressure stimulation: Pressure pain detection thresholds (pPDT) were assessed at C5 (clavicle), T10 dorsal (pancreatic viscerotome), T10 ventral (upper abdomen pancreatic viscerotome), L1 (iliac crest) and L4 (quadriceps) on the dominant side using an electronic pressure algometer with a steadily increasing pressure rate of 30 kPa/s (Algometer Type II, SOMEDIC Electronic, Solna, Sweden). pPDTs were summed to generate an overall pain sensitivity score. A pPDT index, representing segmental pressure hyperalgesia, was calculated by dividing the sum of pPDTs at pancreatic dermatomes (T10 ventral and dorsal) by the sum at reference sites (C5, L1, L4). Cold pressor test: The dominant hand was immersed in ice‐chilled water at 2°C (approximately 36°F) for a maximum of 120 s, with endurance time recorded (seconds). Pain intensity was rated every 10 s on a 0–10 numeric rating scale (NRS). Conditioned pain modulation (CPM): CPM, considered an indirect measure of descending pain modulation, reflecting inhibitory signals from the brainstem that dampen nociceptive input at the spinal cord, was induced using the cold pressor test as a conditioning stimulus and assessed by measuring pPTT at L4 before and after conditioning. The relative change (%) in pPTT was calculated as a measure of CPM efficiency. Temporal summation: Temporal summation, considered a proxy of neuronal excitability, was assessed using an 8 mN pin‐prick device (PinPrick Stimulatoren, MRC Systems GmbH, Heidelberg, Germany). Pain intensity was rated on a 0–10 NRS after a single stimulation as well as after 10 repetitive (1 Hz) stimulations on the ventral pancreatic dermatome and the dominant forearm. The difference in pain intensity scores between the first and tenth stimulus was used as a measure for temporal summation. P‐QST phenotypes were defined based on previously published criteria (Phillips et al.  2020 ). These criteria are based on previously published normative reference values derived from P‐QST examinations in 122 healthy controls, using the 10th/90th percentile thresholds to define abnormality (Phillips et al.  2020 ). Widespread hyperalgesia was defined as an increased pain sensitivity (test result outside the 10th/90th percentiles) for at least two of the following P‐QST parameters: CPM, cold pressor endurance time, pPDT sum and temporal summation at the forearm. Segmental hyperalgesia was defined as an increased pain sensitivity (outside the 10th/90th percentiles) for at least two of the following P‐QST parameters: pPDT index or temporal summation at the ventral dermatome without meeting criteria for widespread hyperalgesia. Patients who did not meet the definition of either widespread or segmental hyperalgesia were characterised as having no hyperalgesia. Changes in P‐QST phenotypes (trajectories) between baseline and follow‐up were classified as: Improving: progression from widespread to segmental or no hyperalgesia, progression from segmental hyperalgesia to no hyperalgesia. Stable: no change in P‐QST phenotype. Deteriorating: progression from no hyperalgesia to segmental or widespread hyperalgesia, progression from segmental to widespread hyperalgesia. Improving: progression from widespread to segmental or no hyperalgesia, progression from segmental hyperalgesia to no hyperalgesia. Stable: no change in P‐QST phenotype. Deteriorating: progression from no hyperalgesia to segmental or widespread hyperalgesia, progression from segmental to widespread hyperalgesia. Trajectories of P‐QST phenotypes were presented using a Sankey diagram. Descriptive statistics for baseline characteristics were reported as means and standard deviations for continuous variables and as counts and percentages for categorical variables. Baseline differences in PROMs across P‐QST subgroups were evaluated using the Kruskal‐Wallis test due to non‐normal score distributions. Dunn's post hoc tests with Holm adjustment were applied to identify pairwise differences when overall group effects were significant. Longitudinal trajectories in P‐QST parameters and PROMs were analysed using linear mixed‐effects models with random intercepts for patients and fixed effects for time. Associations between changes in P‐QST phenotypes and changes in PROMs during follow‐up were assessed using Kruskal–Wallis tests across phenotype groups, followed by Dunn's post hoc tests when relevant. For significant post hoc pairwise comparisons, effect sizes were calculated and reported as rank‐based effect sizes ( r ), computed as the Z ‐statistic from Dunn's test divided by the square root of the total number of observations. Phenotype transitions across all follow‐up visits are illustrated descriptively using a Sankey diagram. However, inferential analyses of phenotype change were restricted to one‐year follow‐up data, as retention rates decreased substantially beyond 1 year, limiting the robustness of multi‐timepoint transition modelling. Missing data were primarily due to loss to follow‐up. No imputation was performed. Linear mixed‐effects models were fitted using maximum likelihood estimation and included all available observations. Analyses requiring paired data were performed as complete‐case analyses, including participants with available data at the relevant time points. All analyses were conducted in R (version 4.4.0) and RStudio (version 2024.04.1). A two‐sided p ‐value < 0.05 was considered statistically significant.

A total of 84 patients were included over a 3‐year period, 49 with CP and 35 with RAP. An inclusion flowchart is shown in Figure  1 . Due to continuous inclusion and loss to follow‐up, the follow‐up period varied between participants: 61 patients (73%) completed 1‐year follow‐up, 42 (50%) completed 2 years and 17 (20%) completed 3 years of follow‐up. Patient characteristics are reported in Table  1 . The mean age was 48.8 ± 16.9 years, 71% were male, 42% had RAP and 58% CP. The most common aetiological risk factor was smoking (55%). Pain status at baseline included 12% with no pain, 71% with intermittent pain and 17% with constant pain. Inclusion and follow‐up flowchart. Flowchart showing the number of patients included along with P‐QST phenotypes. Baseline demographic and clinical characteristics of the cohort. Note: Nutritional risk factor was defined as a diet with high fat and protein intake or dyslipidaemia (elevated triglycerides) (Schneider et al.  2007 ). Abbreviations: CP, chronic pancreatitis; HADS, Hospital Anxiety and Depression Scale; RAP, recurrent acute pancreatitis. To assess potential attrition bias, we compared baseline QST and pain‐related PROMs between participants who completed follow‐up and those who did not. Participants who did not complete the 1‐year follow‐up had lower cold pressor tolerance at baseline (median 37 vs. 120 s, p  < 0.001), higher BPI interference (median 2.0 vs. 0.71, p  = 0.041), lower global quality of life (median 50 vs. 66.7, p  = 0.042) and a higher proportion of widespread hyperalgesia at baseline (40% vs. 14%, p  = 0.032). The change in P‐QST parameters and corresponding P‐QST phenotypes over time was analysed to explore changes in pain sensitisation patterns. Table  2 summarises changes in P‐QST parameters over time. Patients tolerated the cold pressor test for increasingly longer durations over time, with an estimated mean increase of 4.0 s per year ( p  = 0.005). Likewise, the sum of pain detection thresholds rose by an estimated 269 kPa per year ( p  < 0.001). Temporal summation on the forearm increased by an average of 0.21 NRS points annually ( p  = 0.015), while the pPDT index showed a yearly decline of 0.04 ( p  = 0.023). P‐QST parameters over time. Note: Group means for CPM, cold pressor endurance time, pressure pain thresholds and temporal summation across visits. All data are presented as mean (SD). Abbreviations: CET, cold pressor endurance time; CPM, conditioned pain modulation; pPDT, pressure pain detection threshold; TS, temporal summation. P‐QST phenotypes at baseline are shown in Figure  2 . At the initial assessment, 47 patients (56%) had no signs of hyperalgesia, 21 (25%) had segmental hyperalgesia and 16 (19%) had widespread hyperalgesia. Changes in hyperalgesia status from baseline to first follow‐up occurred in 28 patients (46%). Fourteen patients (23%) had improvement in hyperalgesia patterns, and 14 (23%) experienced deterioration (Figure  2 ). QST trajectories. Distribution of QST phenotypes (no hyperalgesia, segmental hyperalgesia and widespread hyperalgesia) at baseline and follow‐up visits. Green symbolises no hyperalgesia, orange symbolises segmental hyperalgesia, blue symbolises widespread hyperalgesia and light blue symbolises lost to follow‐up. At baseline, there were significant differences in PROM scores across P‐QST subgroups, with patients with widespread hyperalgesia reporting lower quality of life and higher pain severity, interference, anxiety and depression scores (all p  < 0.05, Table  3 ). Patient‐reported outcome measures by QST phenotype at baseline. Note: Quality of life, BPI severity and BPI interference scores at baseline visit, stratified by QST phenotype. p ‐values indicate differences between P‐QST subgroups. All data are presented as mean (SD). Abbreviations: BPI, Brief Pain Inventory; QoL, quality of life. Figure  3 shows PROM scores across the study period. Quality of life scores increased over time, with an estimated mean improvement of 2.92 points per year ( p  < 0.001). Pain interference declined, with a mean yearly reduction of 0.24 points ( p  = 0.016), while pain severity showed a trend towards reduction (−0.15 points per year, p  = 0.052). PROM scores over time. Boxplots showing changes in quality of life, BPI severity and BPI interference scores over the study period. BPI, Brief Pain Inventory; EORTC GHS, European Organisation for Research and Treatment of Cancer Global Health Score; QoL, quality of Life. In pairwise Dunn's post hoc comparisons, p atients with increasing hyperalgesia from baseline to 12‐month follow‐up experienced significant reductions in quality of life as compared with those with decreasing hyperalgesia ( r  = 0.295; p  = 0.02; Figure  4 ). A similar pattern was seen for pain interference: patients with worsening hyperalgesia levels had higher BPI interference scores compared to those with stable hyperalgesia levels ( r  = 0.396; p  = 0.018). Pain severity scores trended towards a difference between the improving and worsening hyperalgesia groups ( r  = 0.290; p  = 0.074). Associations between changes in P‐QST phenotypes and PROMs during follow‐up. Box plots illustrating changes in quality of life (A), BPI severity (B) and BPI interference scores (C) from baseline to follow‐up. The X ‐axis depicts patients who improve in QST phenotype (e.g., moving from widespread hyperalgesia to segmental hyperalgesia or no hyperalgesia), those who are stable in QST phenotype, and those who deteriorate. p ‐values indicate differences between groups. QoL, quality of life; QST, quantitative sensory testing.

In this prospective study of patients with RAP and CP, we found that P‐QST phenotypes showed dynamic changes over time. Almost half of the patients changed phenotype during the first year of follow‐up, and these changes were closely linked to fluctuations in patient‐reported scores of pain interference and quality of life. Patients with widespread hyperalgesia at baseline reported worse pain‐related outcomes compared with those with segmental or no hyperalgesia, and patients whose hyperalgesia worsened during follow‐up experienced greater declines in quality of life and a greater increase in pain interference than those who were stable or improving in P‐QST phenotype. These findings suggest that pain sensitisation in RAP and CP is a dynamic process rather than a fixed trait, and this supports the use of repeated mechanistic assessments to improve understanding of pain trajectories and heterogeneity in clinical research. In the longer term, such information may inform patient stratification and the design and timing of mechanism‐based interventions. Our longitudinal findings provide a nuanced picture of how pain sensitivity might evolve in patients with RAP and CP. P‐QST has been widely used in CP research to detect altered pain processing. Cross‐sectional studies have consistently shown that a substantial subset of patients exhibit segmental or widespread hyperalgesia compared with healthy controls or patients with pain‐free CP (Faghih et al.  2022 ; Kuhlmann, Olesen, Grønlund, et al.  2019 ). These findings have led to the interpretation that central sensitisation plays a key role in pancreatitis‐related pain. However, all studies to date have assessed P‐QST at a single time point, leaving it unclear whether these P‐QST phenotypes represent stable traits or dynamic states that evolve in parallel with clinical parameters such as pain experience and life quality, possibly influenced by changes in treatment and interventional procedures. A recent comprehensive QST study across various pancreatic diseases (acute and chronic pancreatitis, autoimmune pancreatitis and pancreatic cancer) reported notable differences in somatosensory profiles between acute and chronic pancreatic conditions, consistent with disease‐specific alterations in pain processing along the continuum from acute to chronic pancreatitis (Göltl et al.  2025 ). In contrast to several previous QST studies describing sensory gain, this study predominantly observed sensory loss. These discrepancies are more likely explained by methodological differences rather than patient characteristics alone. While the study by Göltl et al. primarily applied QST protocols developed for neuropathic pain, our assessment was specifically designed to capture pain processing in pancreatitis, focusing on clinically relevant parameters for visceral and deep pain. Our findings attest to previous studies by showing that P‐QST measures and phenotypes are not fixed. In our cohort, summed pain detection thresholds and cold pressor endurance improved, suggesting a reduction in overall pain sensitivity. In contrast, the pain detection index declined, indicating that segmental and central sensitisation may persist or even intensify in some patients. Individual QST indices showed heterogeneous longitudinal patterns, indicating that not all components of pain processing changed in parallel during follow‐up. Pressure pain detection thresholds and cold pressor tolerance showed systematic changes over time, whereas temporal summation and CPM changed only modestly. This may reflect actual differences in responsiveness across QST domains where dynamic measures such as CPM and temporal summation are less reproducible than static measures (Olesen et al.  2012 ), but could also be influenced by factors such as measurement variability and contextual effects. Clinically, these findings suggest that repeated QST captures changes in some experimental pain responses, but that single indices should not be interpreted as direct proxies of symptom burden. The modest associations between QST indices and patient‐reported outcomes may be explained by the fact that QST assesses experimentally evoked responses under controlled conditions. In contrast, PROMs capture multidimensional clinical outcomes shaped by ongoing nociceptive input, coping, mood and treatment factors. Most longitudinal QST studies in chronic pain patients have been focused on evaluation before and after interventions, limiting our knowledge of changes that can naturally occur over time. There have been studies on longitudinal tendencies in QST measures in patients with low back pain, one reporting decreasing pain tolerance thresholds and stable temporal summation over 6 months in a small cohort of 28 patients with chronic nonspecific neck pain (Ortego et al.  2022 ), and one reporting results where QST results correlated with current pain intensity levels (McPhee and Graven‐Nielsen  2019 ). Our results attest to these findings, with longer follow‐up and a larger sample size, adding new insights by showing that P‐QST measures and phenotypes can fluctuate over several years and do not necessarily deteriorate over time. Indeed, only 22% of our cohort worsened in pain phenotype from baseline to the first‐year follow‐up, and the general increase in pain detection threshold suggests that most patients are stable or improving. These findings suggest that central sensitisation is not a fixed state but may be dynamic, and in some patients, reversible over time, even in the context of standard clinical care (Woolf  2012 ). Patients generally improved in PROMs over the follow‐up period, with better quality of life and lower pain interference. These changes were related to improvements in the P‐QST phenotype. The observed improvements could be related to successful analgesic treatment or improved resilience, where patients adapted to their chronic pain via, for example, improved coping and social support (Farnes et al.  2024 ; Sturgeon and Zautra  2010 ). The observed improvements primarily reflect group‐level trends and mask substantial heterogeneity within the cohort. A general improvement does not preclude that some patients experience deterioration over time. Notably, changes in PROM scores paralleled shifts in P‐QST phenotypes. Patients who developed more sensitised phenotypes were also less likely to show improvement in their PROM scores. This observation is consistent with current literature supporting a bidirectional relationship between QST‐derived sensitisation and patient‐reported outcomes. On one hand, several systematic reviews and meta‐analyses have demonstrated that heightened sensitisation, as measured by QST, is associated with higher pain intensity, greater disability and poorer quality of life across musculoskeletal and chronic pain conditions (Georgopoulos et al.  2019 ; Hübscher et al.  2013 ). Moreover, baseline QST parameters have been shown to predict treatment response and future pain trajectories, suggesting that sensitisation may act as a pathogenic driver of unfavourable outcomes (Arant et al.  2022 ; Georgopoulos et al.  2019 ; Moore et al.  2020 ; Petersen et al.  2021 ). On the other hand, evidence also indicates that this relationship may not be unidirectional. Higher pain burden or longstanding symptoms can themselves facilitate sensitisation processes, resulting in aberrant QST profiles. Importantly, QST abnormalities have been shown to normalise following effective pain interventions in some cohorts, implying that sensitisation may, at least in part, be secondary to persistent nociceptive input or maladaptive pain processing (Arant et al.  2022 ; Kennedy et al.  2021 ; McPhee and Graven‐Nielsen  2019 ). Taken together, these findings underscore a dynamic interplay in which sensitisation and patient‐reported pain outcomes reinforce each other over time. Within this context, our results suggest that patients who progress towards a more sensitised phenotype may enter a self‐perpetuating trajectory of pain amplification and limited symptomatic improvement. The strength of this study involves the design, as it is the first prospective, multicentre, long‐term longitudinal P‐QST investigation in patients with RAP and CP. Furthermore, by integrating validated PROMs with P‐QST measures, the study provides a multidimensional assessment of pain. Several limitations must be acknowledged. First, patients were recruited exclusively from two Danish tertiary pancreatic centres, which may limit external validity given Denmark's relatively homogenous population with limited ethnic diversity (Kærgård  2010 ). Second, P‐QST was used as a proxy for central sensitisation. While it reflects sensory abnormalities, it cannot definitively identify underlying neurological mechanisms (van Driel et al.  2024 ). Third, patient retention was a challenge. Reasons for loss to follow‐up were not systematically recorded, as the study was integrated into routine clinical follow‐up, and missed visits could not always be accounted for. Part of the reduced follow‐up rate beyond one‐year follow‐up was attributable to late inclusion in this ongoing project. These cases were not true dropouts and, therefore, unlikely to bias the results. Fourth, potential habituation effects from repeated P‐QST testing cannot be excluded, although the one‐year interval between examinations likely minimised the risk. Fifth, the effect sizes for differences in PROM changes across P‐QST phenotype groups were small to moderate ( r  = 0.30–0.40). This suggests that phenotype change explains only a limited proportion of the variability in patient‐reported outcomes and that other clinical and psychosocial factors are likely significant contributors. Participants lost to follow‐up had worse baseline pain‐related profiles, suggesting potential attrition bias; therefore, longitudinal estimates should be interpreted cautiously. Finally, the study was not designed to assess the impact of specific treatments on longitudinal changes in P‐QST measures or PROMs. As the patients often received multiple concurrent interventions, including endoscopic procedures and pharmacological pain management, treatment effects could not be isolated. Consequently, observed improvements cannot be attributed to specific therapies. This limitation is being addressed in the ongoing INPAIN study, which is specifically designed to evaluate treatment‐related changes in pain phenotypes in a larger cohort (Hagn‐Meincke et al.  2025 ).

This longitudinal study demonstrates that P‐QST phenotypes in patients with RAP and CP are dynamic and that changes in P‐QST phenotypes are associated with clinically meaningful changes in patient‐reported outcomes. The findings challenge the assumption that pain in pancreatitis inevitably leads to chronification with progressive sensitisation and suggest that sensitisation is, at least in part, dynamic and potentially reversible.

Chronic pancreatitis (CP) is an inflammatory disease of the pancreas, typically characterised by intermittent or chronic abdominal pain (Moran et al.  2015 ). CP often develops as a consequence of recurrent episodes of acute pancreatitis (i.e., recurrent acute pancreatitis; RAP), and thus represents a late stage of a progressive inflammatory process (Cook et al.  2023 ). Pain affects more than 80% of patients during the course of RAP and CP and has a major impact on quality of life (Gardner et al.  2010 ). Although invasive procedures may alleviate pain in selected patients with stones and/or strictures in the pancreatic duct, the presence of pain and its severity often do not correlate with morphological findings (Brøndum Frøkjær et al.  2013 ; Wilcox et al.  2015 ). This mismatch underscores the complex pathophysiology of pancreatitis‐related pain, which involves peripheral and central nervous system changes (Olesen et al.  2010 , 2017 ). Quantitative sensory testing offers a standardised method to assess sensory function, used to characterise pain sensitivity in various pain disorders (Katz et al.  2015 ; Roldan and Abdi  2015 ; Zaslansky and Yarnitsky  1998 ). A pancreatic quantitative sensory testing (P‐QST) protocol has been developed for characterising pain processing in patients with pancreatic diseases (Arendt‐Nielsen and Yarnitsky  2009 ; Kuhlmann, Olesen, Olesen, et al.  2019 ). Based on established normative criteria for sensory parameters, the P‐QST protocol classifies patients into three mutually exclusive groups: no hyperalgesia, segmental hyperalgesia (in the pancreatic viscerotome at the Th10 level), or widespread hyperalgesia, representing increasing levels of neural sensitisation (Phillips et al.  2021 ). It is currently not known how P‐QST parameters and phenotypes change over time. Longitudinal studies have demonstrated that pain in patients with CP is not a static phenomenon but evolves over time. Some patients experience progressive patterns, transitioning from no pain to intermittent or constant pain, whereas others exhibit fluctuating symptoms or improvement (Kempeneers et al.  2020 ). Comparable temporal dynamics have been observed in quality of life among patients with CP, linked to variations in pain intensity and frequency (de Rijk et al.  2023 ). Notably, these changes are often not explained by morphological findings on cross‐sectional imaging, suggesting that alterations in pain processing may play a key role throughout the disease course. While patient‐reported outcome measures (PROMs) capture the subjective burden of pain and its impact on daily life (Han et al.  2024 ; Kuhlmann et al.  2021 , 2022 ), they do not provide insight into sensory processing within the nervous system. P‐QST complements PROMs by identifying sensory abnormalities indicative of altered pain processing (Braun et al.  2021 ; Weaver et al.  2021 ). Together, these approaches offer an understanding of pain in RAP and CP. However, although longitudinal changes in PROMs have been documented in patients with CP, it remains unclear whether P‐QST phenotypes represent stable traits or dynamic states that vary with disease progression or fluctuations in clinical pain. We hypothesised that P‐QST phenotypes are dynamic and change synchronously with PROMs. The aims were to characterise three‐year P‐QST phenotype trajectories in RAP and CP and to assess associations with pain severity, interference and quality of life.

The authors declare no conflicts of interest.

Data S1: ejp70250‐sup‐0001‐supinfo.docx.