Phenotyping peripheral neuropathies with and without a cross-sectional multicenter study

Authors: Panoraia Baka, Daniel Segelcke, Frank Birklein, Esther M. Pogatzki-Zahn, Stephan Bigalke, Ayşenur Süer, Martin Dugas, Livia Steenken, Claudia Sommer, Aikaterini Papagianni

Abstract

Pruritus in peripheral neuropathy is associated with small-fiber dysfunction and pathology and impacts on depression and anxiety levels.

1. Introduction

Chronic pruritus entails a significant burden of disease because of its annoying unpleasantness.^43^ Neuropathic pruritus accounts for approximately 8% of all chronic pruritus cases^57^ and occurs in neurologic disorders of both the central^15,16,27,41,50,59,62,70,72,73^ and the peripheral nervous system.^1,37,44,52,66^ This type of pruritus resulting from various peripheral neuropathies (PNP) has been discussed,^40,42^ but probably because of lack of large-scale data, it remains underrecognized among physicians.^8^ Neuropathic pruritus has received considerably less attention than neuropathic pain and remains poorly characterized.^48,67,69^ For instance, psychophysical studies using quantitative sensory testing (QST) have facilitated the sensory phenotyping of patients and allowed for the identification of distinct subgroups based on their sensory profiles.^2,68^ Subgroups might indicate potential underlying mechanisms and may have the potential to inform on personalized treatment approaches.^2^ Another method to gather information about peripheral nociceptors in PNP is the analysis of intraepidermal nerve fibers, particularly their density (intraepidermal nerve fiber density (IENFD)) in skin punch biopsies. Skin biopsy for IENFD analysis is recommended in the current guidelines on the diagnostics of neuropathic pain.^63^

Both neuropathic pruritus and pain might share common elements and overlapping mechanisms.^6,7,19,24,29,38,51^ However, there is a need for comprehensive investigations into potential phenotype differences among patients with PNP with and without pruritus, as well as an exploration of the resulting effects on patients' functionality and quality of life (QoL). Meticulous phenotyping is vital for accurately defining the characteristics of individuals who experience pruritus and contrasting them with those without.^61^ Such an approach holds the potential to yield valuable insights into the underlying mechanisms of neuropathic pruritus, discern disparities between the mechanisms governing pain and pruritus, and contributing to the research on potential therapeutic options for patients coping with chronic pruritus.

In this multicenter study, we characterized the phenotype of patients with PNP with and without pruritus and pain, through a comprehensive neurological examination, assessment of nerve conduction studies (NCS), QST, and skin biopsy with determination of IENFD. Patient self-reported questionnaires were used to assess the impact of both pruritus and pain on patients' daily functioning and QoL.

2. Materials and methods

2.1. Study design

We conducted a cross-sectional multicenter study in 3 university hospitals in Germany (Departments of Neurology of the University Hospitals Mainz and Würzburg, Department of Anesthesiology of the University Hospital Münster). The study protocol adhered to principles of the Helsinki Declaration and was approved by the local ethics committees of all 3 study centers (Mainz 837.437.17, Würzburg 239/17, Münster 2017-576-f-S). All study participants provided written informed consent before inclusion.

We performed standardized procedures in all 3 study centers. Skin biopsy analysis was performed in one reference center (Würzburg). All group members of the 3 teams met at least twice a year during the entire recruitment period (2019-2021) to monitor study protocol adherence.

Adult ambulatory patients (≥18 years old) with PNP of any etiology were included. The study included patients who were referred to our outpatient departments, predominantly seeking a second opinion. These patients had either a confirmed or suspected diagnosis of PNP, irrespective of the etiology. To be eligible for enrollment, they underwent a screening to assess the presence of pain or pruritus symptoms. Exclusion criteria encompassed significant psychiatric disorders, acute skin infections, other forms of chronic pruritus (such as dermatologic, systemic, psychogenic, or drug-related), or chronic pain conditions (including chronic low back pain, painful osteoarthritis, and headaches). In addition, individuals using oral anticoagulants (unless they ceased usage before skin punch biopsies) or those unable to provide informed consent were excluded. A control group was established comprising healthy study participants recruited through advertising. Participants were excluded from the control group if they met any of the following signs or symptoms of PNP, chronic or current pain or pruritus, psychiatric or neurological diseases, malignant conditions, dermatological diseases as well as any factor that might cause PNP, including diabetes mellitus, pathological glucose tolerance, harmful alcohol use, present or past chemotherapy, or a positive family history of PNP.

Demographic data, general medical and drug history, and symptom duration were obtained together with a detailed history about presence, localization, intensity of sensory symptoms focusing on neuropathic pain and pruritus. Mean and worst intensity of pain and pruritus were quantified using an 11-point (0-10) numerical rating scale (NRS). After filling in questionnaires about affective symptoms and QoL, the study participants underwent a structured neurological examination. This examination allowed calculation of the modified Toronto clinical neuropathy scale (mTCNS)^9^ and the overall disability sum score.^33,34^ Neurological examination was followed by a set of technical NCS, QST, and skin biopsy for determination of IENFD. Based on the results of clinical examination, NCS, QST, and histological findings in skin biopsies, diagnosis of either large-fiber (or mixed) or small-fiber PNP were established. For small-fiber neuropathy (SFN), we performed a multimodal approach involving clinical assessment, QST findings, and IENFD, in accordance with the diagnostic criteria by Devigili et al.^11^

According to the reported mean pain and pruritus intensity during the 4 weeks before inclusion, patients were subdivided into 4 “pruritus,” “pain,” “pruritus and pain,” and “no pruritus/no pain.” Definition of the subgroups was as “pruritus” = NRS ≥3 for itch and <3 for pain, “pain” = NRS ≥3 for pain and <3 for itch, “pruritus and pain” = NRS ≥3 for both itch and pain, and “no pruritus/no pain” = NRS <3 for itch and pain. A cut-off of NRS 3 was chosen because NRS 3 is often used as an inclusion threshold for intervention trials.^4,22^ The study design is presented in the flowchart diagram (Fig. 1). Patients were examined while in a stable state of analgesic or, if present, antipruritic medication.

2.2. Questionnaires

Participants completed a detailed tablet-based set of questionnaires consisting of the Hospital Anxiety and Depression Scale (HADS),^21,53^ the Short Form-12 questionnaire (SF-12),^13,17^ as well as the State Trait Anxiety Inventory (STAI) Form X1 and X2.^30^ The NeuroDerm questionnaire (NeuroDerm/ Arbeitsgemeinschaft Pruritus Forschung [AGP])^71^ and the QoL with pruritus 5PLQ questionnaire^74,75^ were used to evaluate properties of pruritus like quality, intensity, localization, and interference with daily life. Pain-related aspects were investigated by the Brief Pain Inventory,^46,60^ the Neuropathic Pain Symptom Inventory,^5,54^ and the Pain Catastrophizing Scale.^36,58^ German language validated versions of all questionnaires were used.

2.3. Nerve conduction studies

Motor NCS of ulnar and tibial nerves and sensory NCS of ulnar and sural nerves were performed unilaterally (body side was determined by clinical findings) in a warm room using standard methods.^26^ In accordance with European Academy of Neurology -Peripheral Nerve Society (EAN/PNS) criteria,^65^ axonal damage was defined as reduction of compound muscle action potential baseline to first negative peak amplitude, or of sensory nerve action potential, below the lower limits of normal reference values.

2.4. Quantitative sensory testing

Quantitative sensory testing was performed unilaterally according to protocol of the German Research Network of Neuropathic Pain (DFNS)^49^ at the distal lateral leg (test area) and on the ipsilateral cheek (control area). Thermal and mechanical detection, pain thresholds, paradoxical heat sensation, dynamic mechanical allodynia, wind-up ratio, and vibration detection threshold were determined.

2.5. Skin biopsies and determination of intraepidermal nerve fiber density

Two skin punch biopsies were obtained according to standard methods,^28^ from the same body side tested with NCS and QST, 10 cm proximal to the lateral malleolus (lower leg) and 20 cm distal to the spina iliaca (upper leg) with a disposable 6-mm punch under local anesthesia. Skin biopsies exceeding the standard 3 mm size typically used for diagnostic purposes were obtained for comprehensive histological examination in follow-up projects unrelated to the objectives of this study and for long-term preservation in the biobank for future analyses. All skin samples were processed to assess IENFD and the presence of inflammatory changes according to a previously published protocol.^64^ Skin biopsies were analyzed in one center (Würzburg). Intraepidermal nerve fiber density was determined following standardized counting rules^28^ by an investigator blinded to subject allocation.

2.6. Data analysis plan

Statistical analyzes were performed using IBM SPSS 23 Statistics, version 23.0, GraphPad Prism 9 for Windows, and OriginPro 2023. The level of statistical significance was set at P < 0.05. Data comparison was performed using Mann Whitney U test, nonparametric analysis of variance, and respective post hoc analyzes. Categorical data were analyzed with the χ^2^ test. Spearman Rho correlation analyzes were performed to explore associations between clinical, histological, and neurophysiological data. To evaluate QST parameters without being influenced by their physical dimensions, a z-transformation was applied to the QST data. These transformed data were then compared with data from controls matched by age and gender, using the mean of the control data (where z = 0) as the established population mean, as evaluated before.^49^ Principal component analysis (PCA) was conducted on multivariate datasets after standardization for psychological profiling, sensory profiling, IENFD, and multilevel profiling of patients with PNP. The 2 principal components (PC1 and PC2) were developed by identifying the directions (arrow direction) of maximum variance in the data, based on the largest eigenvalues of the covariance matrix. The bi-plots was used to graphically represent the sample scores on the PCs and the loadings of the variables as vectors (arrows), displayed in the same dimension as the cluster plots. All variables incorporated into the PCA are displayed in the bi-plots, illustrating the contribution of each variable to the PCs and the distribution of the samples in relation to these components. Although groups were incorporated into these bi-plots for visualization, they were not factored into the PCA process. The importance of group differentiation was assessed using a multivariate analysis of PC loadings, coupled with Tukey post hoc tests. Further analysis used multivariate analysis of variance. Multivariable multinomial logistic regression was used to identify variables linked to the group classification. No data imputation was performed for missing data. Analyzes were performed by the stepwise and randomized integration of the different data sets into the model. The models were not evaluated regarding gender and age.

3. Results

3.1. Demographics and clinical characteristics

We included 191 patients with PNP {104 females, 87 males, median age (95% confidence interval [CI]), 57 (54-58)} and 57 control subjects (29 women and 28 men, median age [95% CI], 53 [23-76]) (Fig. 1). Of these patients, 138 had SFN and the others had large-fiber or mixed-fiber PNP of diverse origin (see Supplementary Table 1, http://links.lww.com/PAIN/C75). Twenty-five patients (28%) with abnormal NCS, suffering from neuropathic pain and having a reduction or loss of IENFD in skin punch biopsies, were defined having a mixed-fiber PNP. No differences in age between large-, mixed-, or small-fiber PNP were observed. Patients diagnosed with mixed-fiber PNP exhibited a prolonged duration of symptoms (median: 72 months) in comparison with those with large-fiber PNP (median: 24 months) or SFN (median: 36 months).

The 4 subgroups were formed by 21 patients with “pruritus” (9 women, median age [95% CI], 66 [31-81] years), 86 patients with “pain” (50 women, median age [95% CI], 57 [20-81] years), 66 patients with both “pruritus and pain” [36 women, median age [95% CI], 55 [26-85] years) and 18 patients without pain nor pruritus “no pruritus/no pain” (9 women, median age [95% CI], 57.5 [24-77] years). In total, 45.6% of the patients experienced pruritus, either on its own (11%) or in combination with pain (34.6%). No difference was observed regarding sex distribution between groups. The “pruritus” group was older than the control group (P = 0.002) and the “pruritus and pain” group (P = 0.019) (Table 1). History of hypertension (19%), thyroid gland dysfunction (8%), and spinal disk herniation (8%) were most common in our patients. Neurological impairment, as assessed by the mTCNS was different between the groups. Patients in the “pruritus” group were less affected (vs “pain” P= 0.006; vs “pruritus and pain” P < 0.000). Overall disability in daily life, according to the overall disability sum score (ODSS) did not differ between patient groups (Table 1).

3.2. Pruritus characteristics

Median intensity of pruritus did not differ between the “pruritus” and “pruritus and pain” group (median [min-max], 5 [3-10] and 4 [0-8] on NRS, P = 0.069). The following data from the NeuroDerm questionnaire refer to all patients reporting pruritus ≥3. Pruritus lasted 1 to 10 years in 57% and 10 years or longer in 24% (Table 1). Most patients with pruritus (71%) experienced pruritus as “attack-like” episodes, occurring throughout the day in 55.7%. When asked about their emotional experience relating to pruritus, the most common descriptions patients provided were “intractable” (38.5%), “excruciating” (33.7%), and “malicious” (33.7%), whereas some patients expressed that pruritus made them feel “depressed” (29.8%). Factors such as scratching (41%), rubbing (35%), bed warmth (31%), sweating (30%), resting (28%), physical activity (25%), and emotional stress (23%) increased neuropathic pruritus intensity. Patients described their pruritus as “persisting” (39%), “intensely painful” (34%), and “emotionally distressing” (30%). The distribution of pruritus and pain was different. Pruritus was mainly reported in the lower legs and trunk (Fig. 2A), whereas pain was most often reported in the feet, lower and upper legs, and hands, therefore following a length-dependent pattern in the majority of patients (Fig. 2B). Patients with mixed-fiber PNP and SFN reported higher itch intensities on examination day when compared with patients with large-fiber PNP (NRS median itch intensity 5 points, 4 points, and 1 point, respectively).

3.3. Anxiety, depression, and quality of life

As expected, patients with PNP scored higher than healthy control (HC) in HADS, STAI, and SF-12 (P < 0.005 for all tests, Mann–Whitney U test). Despite higher anxiety and depression scores in the HADS, all patients, including those with “pruritus,” remained below the normal cut-off (cut-off ≤8 points;^56^).

A targeted analysis revealed that the “pruritus” group scored higher than HC for depression (P = 0.029) but not for anxiety. By contrast, the “pain” and the “pruritus and pain” groups scored higher in HADS for both anxiety (P < 0.005 for both groups) and depression (P < 0.005 for both groups), also compared with HC. In addition, the “pruritus and pain” group had a significantly higher score in HADS depression compared with “no pruritus/no pain” group (P < 0.05). An additional assessment of anxiety was conducted using the STAI-X1 and STAI-X2 instruments. Here, anxiety levels in the “pain” and “pruritus and pain” groups only were higher than in controls (P < 0.0001 for both).

In the physical aspects of QoL as measured by SF-12, both the “pruritus and pain” group and the “pain” group had significantly lower scores compared with the “pruritus” group (both P < 0.001), the “no pruritus/no pain” group (P = 0.016 and P = 0.048), and healthy control group (both P < 0.001) (Table 2 for details). In the mental QoL-related aspects, the “pruritus and pain” and “pain” groups scored lower than the “no pruritus/no pain” group (P = 0.021, P < 0.05, respectively) and healthy control group (both P < 0.001).

The multivariate analysis confirmed the univariate approach. Using multivariate analysis to synthesize data from the questionnaires on anxiety, depression, and QoL, a distinct phenotype was identified in both the “pain” (P < 0.001) and the “pruritus and pain” (P < 0.001) groups compared with the control group (Fig. 3A). This phenotype was characterized by elevated anxiety (HADS, STAI, and positive loadings at PC1 axis), heightened depression (HADS positive loadings at PC1 axis), and a diminished QoL (SF-12, negative loadings at PC1 axis), especially prevalent in patients exhibiting pain as a primary symptom (Fig. 3B). Notably, these findings starkly contrasted with the “pruritus” group, which did not show this “psychological” phenotype.

3.4. Nerve conduction studies

A total of 167 patients and 47 control subjects underwent NCS as outlined in Supplementary Table 2, http://links.lww.com/PAIN/C75. In general, the values of the patient groups were lower than the HCs. In post hoc analyses, there was no conclusive difference between the different patient groups, nor between the “pruritus,” “pain and pruritus” and “pain” groups vs HC. It was only observed that the “no pruritus/no pain” group exhibited a more severe axonal large-fiber damage with significantly lower sensory sural amplitude when compared with “pain” and to “pruritus and pain” (P = 0.001 and P = 0.002 respectively). For details, see Supplementary Table 2, http://links.lww.com/PAIN/C75.

3.5. Quantitative sensory testing

All patient groups showed significantly altered thermal sensory limen vs HC. In addition, there was a significant increase in cold detection threshold (CDT) in “pruritus” and “pruritus and pain” patients, along with a trend of heightened warm detection threshold compared with HC (Fig. 4A and Table 3). The univariate findings illustrate a sensory phenotype marked by a diminished sensitivity to cold and warm, pointing to functional deficiencies in the Aδ and C fibers and thermal hypesthesia. This is corroborated by significant group segregation in the multivariate analysis, evident in all patient groups except for the “no pruritus/no pain” group (Fig. 4B). An increase of the vibration detection threshold was observed across all patient groups, along with a reduction in the mechanical detection threshold, particularly in the “pruritus and pain” group (Fig. 4C and Table 3). This aspect, highlighted in multivariate analysis, revealed a notable group segregation in the “pain” and “pruritus and pain” patients when pitted against the HC, suggesting a functional impairment in the cutaneous Aß-fiber network, which can be classified as mechanical hypesthesia (Fig. 4D).

3.6. Determination of intraepidermal nerve fiber density

A total of 195 skin punch biopsy samples were taken from the thighs and 215 from the lower legs of patients with PNP and HC without adverse events or complications. Of these, 193 participants had biopsies from both locations (Supplementary Table 3, http://links.lww.com/PAIN/C75). The samples from the “pruritus” group exhibited a reduced IENFD at the proximal biopsy site compared with the “pain” group (P = 0.013) and HC (P < 0.05), which remained significant after correction for age (Fig. 5). No significant relationship was identified between the severity of pruritus and IENFD (P = 0.058 for thigh and P = 0.59 for lower leg). Using multivariate analysis, a unique and significant phenotype was evident in the “pruritus” group, which showed a marked reduction in both distal and proximal IENFD when juxtaposed with controls.

3.7. Model development for stratification of patients with peripheral neuropathy

In the stratification of patients with PNP, a model was developed using cross-sectional datasets and multivariate analysis (multinominal logistic regression analysis), selecting one factor from each assessment level. The best model performance was achieved by incorporating SF-12 physical, sural nerve sensory nerve action potential, CDT, mechanical pain threshold, and proximal IENFD ([Akaike-Information-Criterion] AIC = 171, R^2^ = 0.314) (Fig. 6A and Supplementary Table 4, http://links.lww.com/PAIN/C75). This model leads to a data-driven significant group segregation based on the neurological examination, self-reported questionnaires, assessments of motor and sensory nerve conduction between the “pruritus” and “pain” groups (P < 0.001), as well as “pruritus” and “pruritus and pain” groups (P < 0.001). However, it does not distinguish between the “pruritus” and “no pruritus/no pain” cluster. The 2 pivotal variables in the model are SF-12 (mental health) and CDT (Supplementary Table 5, http://links.lww.com/PAIN/C75). The “pruritus” group was characterized by unaffected QoL (SF-12 physical), tendency to mechanical hyperalgesia, cold hypesthesia, and diminished proximal IENFD (Fig. 6B). By contrast, “pain” group was marked by diminished QoL, impaired mechanical pain sensitivity, preserved cold detection, and unaffected IENFD (Fig. 6B). The “pruritus and pain” group exhibited reduced QoL, tendency to mechanical hyperalgesia, cold hypesthesia, and decreased IENFD (Fig. 6B). In comparison, the “no pain/no pruritus” group displayed normal QoL, tendency to mechanical hyperalgesia, cold hypesthesia, and normal IENFD (Fig. 6B). Integrating of cross-sectional datasets facilitates deep phenotyping of patients with PNP, independent of the predominant symptoms. This method results in significant group segregation (Fig. 6C), elucidating distinct subgroups within the broader PNP patient population.

4. Discussion

We deeply phenotyped a large group of patients with PNP, focusing on the symptoms “pruritus” and “pain.” We found several similarities but also differences between patients with pruritus or pain, particularly after focusing on multidimensional data integration. Our key finding is that pruritus is highly prevalent in patients with PNP, in our cohort affecting almost half of patients. It could present as the only neuropathic symptom or in combination with pain, which then did not influence the severity of pruritus. Among the tests performed, we did not find a single biomarker, which would be indicative for neuropathic pruritus. However, neuropathic pruritus coincides with a phenotype pattern that includes an impaired small-fiber function and structure showed by reduced cold perception and diminished IENFD at the thigh, and coincident with normal or heightened mechanical pain perception.

4.1. Pruritus is prevalent in peripheral neuropathy

Patients with PNP, particularly SFN, have been associated with the presence of neuropathic pruritus before,^8,42^ but the prevalence of pruritus in PNP remained unexplored. In our cohort, 11% of patients reported pruritus as their leading symptom, and more than a third of our patients (34.6%) reported both pruritus and pain, both in a significant strength. Less than 10% of our patients reported none of these symptoms, which is lower for the symptom of pain compared with previous reports (approximately 24%).^76^ Previous reports stating percentages of neuropathies with or without pruritus were not found. We had excluded concomitant causes of itch/pruritus (dermatologic, systemic, psychogenic, or drug-related), thus our systematic assessment for pruritus has brought to light the existence of neuropathic pruritus alone or in combination with neuropathic pain in patients with PNP.

4.2. Neuropathic pruritus is often overlooked

Our results suggest that a systematic assessment of pruritus should be performed in patients with PNP. In widely used clinical scales like the mTCNS,^9,42^ which takes patients' symptoms into account, pruritus has not been included as a neuropathic symptom so far. This fact leads to a misleadingly low score for patients with itching as the main symptom, as in our “pruritus” group, falsely suggesting a less severe form of neuropathy than if pain were present. Our results therefore are a plea for a future revision of clinical neuropathy assessment scales. Questions or items related to pruritus should be included to get a more accurate evaluation of patients with neuropathic pruritus.

4.3. …and its distribution differs from pain

Our patients reported pruritus mainly at the lower legs and the trunk. This distribution of itch confirms data from a previous study with neuropathic pruritus, which was associated with SFN. In that study, neuropathic pruritus was observed in the lower extremities, showing a gradient from distal to proximal locations, and it was also reported in the back.^8^ The distribution of pruritus differs from that of neuropathic pain in peripheral neuropathies where the “glove and stocking” distribution is most commonly reported.^23^ Although we have some evidence suggesting that pain originating from glabrous skin is less effectively controlled^12^ and might be therefore predominating in neuropathy, there is only recent evidence showing that different mechanisms exist for hairy and glabrous skin itch. MrgprA3^+^ and MrgprD^+^ neurons mainly mediate hairy skin itch, whereas MrgprC11^+^ neurons are key mediators for glabrous skin itch.^55^ We speculate that the distribution of pruritus in PNP might be explained by a predominant affection of MrgprA3^+^ and MrgprD^+^ neurons coming from neuropathy in proximal hairy skin. Accordingly, we found a reduced proximal IENFD (at the thigh) in our “pruritus” patients. In a previous study, also a non–length-dependent loss or dysfunction of epidermal C fibers has been discussed in relation to neuropathic pruritus.^18^ The lower IENFD of the “pruritus” group at the proximal biopsy site compared with the “pain” group remained significant after correction for age. That is, pruritus in PNP has a non–length-dependent pattern that might correlate to the proximal and less length-dependent small-fiber pathology in these patients. Thus, our study supports the idea that itching in seemingly unaffected skin should trigger further investigations of the skin innervation in these areas in cases of neuropathy.

4.4. Pruritus does not affect quality of life but was associated with depressive symptoms

The patients with PNP in our study did not have significant anxiety and depression according to established cut-off score values. However, as expected and reported before,^20,35,39^ neuropathic pain was associated with higher scores for anxiety and depression and reduced scores for QoL in patient-reported outcome measures. Pruritus, by contrast, was only associated with higher depression scores; the anxiety and the SF-12 outcomes were at the levels of HCs or of patients without neuropathic symptoms. This corroborates the results of a study on neuropathic pruritus in small-fiber neuropathies^42^ but is partially in contrast to a study that was performed on diabetes and pruritus.^45^ These authors described an association of pruritus with neuropathy as in our study, but also with anxiety, depression, and reduced QoL. However, no differentiation was made between different neuropathic symptoms, like pruritus and pain. Therefore, it remained obscure whether the associations in that study were related to pruritus or to PNP itself. Our results suggest that in PNP it is the pain, which is closer associated with affective symptoms than pruritus. In chronic pruritus of other origin, particularly when associated with skin lesions, affective symptoms might be more evident.^14,74^

4.5. Sensory phenotype

Quantitative sensory testing helps to describe somatosensory profiles of patients with PNP and to identify which fiber classes might be functionally most affected. The “pruritus” group in our study was best described as hypesthesia for cold and preserved mechanical pain sensitivity. This sensory phenotype is not significantly different from the patients with pain in our study. However, the preserved mechanical pain sensitivity despite a more advanced Aδ- and C-fiber loss, which is indicated by low CDT and low IENFD in the “pruritus” patients, seems remarkable and is different from what would have been expected in more advanced neuropathies. In these patients, mechanical pain sensitivity usually decreases in parallel to cold perception.^47^ Future studies with more patients and profounder nerve fiber characterization will show whether this translates into a particular pathophysiology of neuropathic pruritus.

4.6. The benefit of cross-sectional datasets integration for deep phenotyping

Integrating cross-sectional patient-level variables is crucial because it enables the concurrent analysis and understanding of a wide array of data types. It can effectively reduce the complexity of these datasets, enhancing their interpretability while simultaneously ensuring minimal loss of critical information.^25^ In contrast to the exclusive analysis of univariate results, researchers and clinicians can acquire a more comprehensive understanding of the complex interactions between patient-level variables. This approach is particularly effective in identifying specific or nuanced phenotypes (cluster analysis),^3^ which may be influenced by an integration of, eg, biological factors, psychological states, and social contexts. Cross-sectional datasets integration in pain or itch^31^ recognizes that health and disease conditions are not solely the product of isolated biological processes (eg, for pain or itch) but are also shaped by psychological, social, and cultural influences. In this study, biological findings such as IENFD and sensory phenotype were integrated with psychological parameters and social context to investigate the multidimensional pathomechanisms in chronic pain or chronic itch in patients with PNP using complex, integrated phenotype descriptions of patient subgroups. Similar to the widely recognized complexity regarding the underlying mechanisms of neuropathic pain, involving not only peripheral generators but also central mechanisms like sensitization,^10,32^ a central–peripheral nervous system interaction may also underlie neuropathic itch. This could potentially explain the presence of symptoms beyond anatomical areas with reduced IENFD, supporting the decision of a multifaceted approach such in our study.

4.7. Limitations

The distinction between individuals experiencing pruritus and those without pruritus, as well as between those with pain and those without, was made with a degree of subjectivity. In our multicenter study, which does not use a population-based epidemiological enrollment, focusing on the presence of pruritus or pain may inherently introduce a recruitment bias. Therefore, we cannot draw definitive general conclusions regarding the prevalence of neuropathic pruritus in patients with PNP. Although the study had a significant number of participants overall, the sample size for specific subgroups was comparatively limited. It is important to acknowledge that the differences observed among the 4 subgroups may be attributed to the combined effects of multiple testing in the statistical analyses and the sample size. Furthermore, the study lacked histochemical or functional characterization of the fibers associated with the observed symptoms, leaving the underlying mechanisms somewhat unclear.

In conclusion, our study indicates a higher presence of neuropathic pruritus in patients with PNP than currently assumed. Given the functional impairment and emotional distress of patients with pruritus, timely adjustments of clinical management are a necessity. Future studies should dive deeper into the molecular mechanisms relevant for pruritus distinct from pain and identify potential biomarkers that can be used for target-specific treatment options in the future. Because many or most of the patients with pruritus are referred to dermatologists, an interdisciplinary diagnostic and therapeutic approach is warranted.

Conflict of interest statement

The authors have no conflicts of interest in relation to this work to declare.

Appendix A. Supplemental digital content

Supplementary Material