Background

Primary biliary cholangitis (PBC) is a chronic cholestatic liver disease with debilitating symptoms, including pruritus and fatigue. Recent large clinical trials have found a baseline prevalence of pruritus in patients with PBC of about 50–70% [1, 2]. Other studies from the UK have shown a significant impact of these symptoms on quality of life [3, 4]. However, an understanding of the impact of pruritus in PBC in the real world, particularly within the USA, is lacking. Multiple potential therapies for cholestatic pruritus have been studied and found to be partially effective, including cholestyramine, rifampicin, naltrexone, and sertraline, but none are FDA-approved for use to treat pruritus in PBC patients specifically. The American Association for the Study of Liver Disease (AASLD) and European Association for the Study of the Liver (EASL) recommend a step-wise approach to treat pruritus in PBC,[5, 6], but the extent of uptake of these recommendations in the medical community is also unknown. This study characterizes the population with pruritus in the TARGET-PBC cohort. The purpose is to describe the population characteristics, the impact of pruritus on quality of life (QOL), and the management practice of pruritus in a real-world cohort.

Patient Cohort

TARGET-PBC is a longitudinal observational cohort of patients with PBC receiving usual care in hepatology or gastroenterology clinics at one of 38 academic and community sites across the USA. The design and a description of the cohort have been presented in more detail elsewhere [7]. Patients with a diagnosis of PBC by a treating physician were eligible for inclusion in the study. Individuals actively enrolled in a clinical trial were excluded.

Methods

Following informed consent, patient data were obtained from both medical records and patient self-report. Medical records for each patient included three years prior to the date of consent and up to five years prospectively. Data were abstracted from the medical records, including clinical notes, laboratory data, medication lists, all prior imaging reports, radiographic and other diagnostic procedures, and all prior liver biopsy reports. Missing data were minimized by performing site queries. In addition to clinical and treatment data, patients were asked to complete patient-reported outcome (PRO) surveys approximately every six months. These PRO surveys included the PBC-40, 5-D Itch, and the PROMIS fatigue survey, described below. The concepts covered by the PRO questions are summarized in Table 1.

Table 1 Patient reported outcome (PRO) survey descriptions
Full size table

Target RWE is the sponsor of TARGET-PBC and is responsible for the data and quality control activities. Data are abstracted from complete medical records which are uploaded into the database by sites for enrolled participants. There are various processes in place to ensure the quality of data collected for the TARGET-PBC study. Edit checks, auto-coding of adverse events and concomitant medications, expert adjudication, and source-document verification are all components of the data quality system.

The analysis reported here focuses on the population who had completed PROs at least once and uses the most recent PRO, clinical and laboratory data. The medication list was developed as of the last medical record abstraction and includes all recorded medications that a patient had been prescribed for PBC and associated cholestatic pruritus.

Patient Reported Outcome (PRO) Surveys

The PBC-40 consists of 40 questions across six domains of interest related to PBC: general symptoms, itch, fatigue, cognition, social, and emotional [8]. Items are scored from zero or one to five and individual item scores are combined to give a total domain score. This questionnaire assesses symptoms over the last 4 weeks. Using the threshold for clinical significance suggested by the developers, clinically significant (CS) itch was defined as ≥ 7 points from a maximum of 15 on the itch domain and mild itch as ≥ 1 and < 7.

The 5-D Itch scale comprises five domains: duration, degree, direction, disability, and distribution [9], with each domain accounting for 5 points. The domain scores are then added together for a total 5-D score, potentially ranging from 5 (no pruritus) to 25 (most severe pruritus). This survey assesses itch over the last 14 days.

The PROMIS fatigue survey (PROMIS Item Bank v1.0-Short Form 8a) evaluates symptoms ranging from a mild subjective tiredness to an overwhelming, debilitating, and sustained sense of exhaustion [10]. The domains include fatigue (frequency, duration, and intensity) and the impact of fatigue (on physical, mental, and social activities). This survey asks patients to rate average fatigue over the past 7 days using a five-point Likert scale from 1 to 5, where 1 means the least impact/severity, and 5 being the most. The total scores for all items are added [total of 8 to 40] and cross-referenced with a lookup table to obtain a T-Score.

Statistical Methods

Descriptive statistics were reported for continuous and categorical variables overall and by itch severity. Continuous variables were summarized using the frequency, median, minimum, maximum, and interquartile range (IQR) values. Categorical variables were summarized using the frequency and the percentage relative to those with non-missing values. Kruskal–Wallis tests were used to compare median differences in symptoms between the mild and CS itch groups (no itch was excluded). Patient characteristics, disease severity, and treatment patterns were compared according to the presence and severity of itching. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Subject Characteristics

A total of 211 out of 671 PBC patients completed PROs allowing the presence (or absence) and severity of itch to be assessed and were included in the current study. Table 2 shows the characteristics of patients who responded compared to those who opted not to complete the PROs. No obvious demographic differences were observed between these groups except that respondents were more likely to be white/non-Hispanic/Latino (p = 0.01) and have a lower GLOBE score (p < 0.05). PBC patients were of a similar age at the time of the survey (respondents–61; non-respondents–63) and had been diagnosed with PBC a similar amount of time (respondents–7.2 years; non-respondents -5.6 years). Patients were predominantly female (respondents–92%; non-respondents–91%), white (respondents–87%; non-respondents–79%), and non-Hispanic (respondents–84%; non-respondents–75%). Patients frequently had Antimitochondrial antibody (AMA) positivity documented (respondents–87%; non-respondents–81%) and had undergone a liver biopsy (respondents–66%; non-respondents–60%). Under half of patients had cirrhosis (respondents–35%; non-respondents –41%). Of the 211 patients within the study, 83% received care from an academic site, while the remaining patients received care at a community site.

Table 2 Demographics of PRO survey respondents vs. non-respondents
Full size table

PBC-40

Itch Domain

The presence of itching of any degree was reported in 170 (81%) patients. The majority of these, 107 (63%) had a mild itch, and 63 (37%) were classified as having a clinically significant itch with a score ≥ 7. Patients with CS itch were younger (CS = 58 y/o; mild itch = 64 y/o; no itch = 64 y/o), more frequently had cirrhosis (CS = 48% vs. mild itch = 27%, and no itch = 37% p = 0.03) and had higher alkaline phosphatase levels (CS = 177 IU/L vs. mild itch = 143 IU/L and no itch = 153 IU/L, p = 0.002) compared to those with mild or no itch, respectively (Table 3). P-values in Table 3 are tests for any difference between the three groups (CS itch, mild itch, and no itch).

Table 3 Patient Characteristics by PBC-40 Itch Severity
Full size table

Other Domains

Across all domains of the PBC-40, those with CS itch scored significantly worse than those with mild itch. There was no notable difference in scores between those with mild itch and those with no itch. The largest difference was seen in cognitive and social domains; median scores in the CS itch group were ~ 80% higher than those in the no itch group, indicating more distress in patients with CS itch. In other domains—fatigue, symptoms, and emotional—the difference was smaller, though still substantial, with median scores 42%, 46%, and 50% greater, respectively, for the CS itch group (Fig. 1).

Fig. 1
figure 1

Median and IQR PBC-40 domain scores by itch severity

Full size image

5-D Itch

The scores for the 5-D Itch Scale were consistent with the PBC-40 itch domain. Respondents with CS itch scored significantly higher (worse) than those with mild itch across all domains (Fig. 2). The direction domain assesses whether itching has gotten better/worse, and both groups scored similarly. Patients with CS itch had a mean duration of itch of less than 6 h per day (although ~ 20% of those with CS itch reported experiencing more than 12 h itching per day) that involved an average of six to 10 body parts, and patients with CS itch reported more widespread itch than those with mild itch, ~ 70% reporting itch affecting > 6 body parts compared with < 20% of those with mild itch. The most commonly reported body parts affected by itch were: head/scalp 67%, lower legs 63%, back 62%, palms of hands 43%, and soles of feet 35%; (data not shown). The majority reported an unchanged itch severity that for most is unchanged over the previous 2 weeks, with 19% reporting a worsening itch. Itch caused significant disability predominantly in sleep (88%), but also occasionally impacting patients’ social life (58%), housework/errands (53%), and work/school (44%) (data not shown).

Fig. 2
figure 2

5-D Itch domains by PBC-40itch severity

Full size image

PROMIS Fatigue

Patients with CS itch reported significantly greater fatigue on the PROMIS fatigue instrument than those with mild and no itch. Individuals with CS itch reported the highest level of fatigue on the following items with a median score of four: “worn out, so tired I had to force myself to do things I needed to do, if I was busy one day I needed at least another day to recover, and I had to pace myself for day-to-day things.” Median scores in the CS, mild and no itch groups were 61, 50 and 50, respectively (p < 0.0001 CS vs. Mild Itch) (Fig. 3).

Fig. 3
figure 3

Mediand and IQR PROMIS fatigue by PBC-40 itch severity

Full size image

Treatments for Pruritus

Overall, patients suffering from CS itch were more likely to receive treatment for itch than those with mild itch (51 vs 28%) (Table 4). However, based on their medical records, 33% of patients reporting CS itch had never received any treatment for itch. These same patients suffering from CS itch were more likely to currently have multiple treatments concomitantly for their underlying PBC documented in their medical record than those with mild itch (32 vs 22%) and to be taking fenofibrate (16%vs 1%). Nearly all patients (97%), regardless of severity of itch, were currently taking ursodeoxycholic acid (UDCA), either alone, or in combination with another medication, while only 16% had received OCA (as a combination or alone) (Fig. 4). Of those receiving pruritus treatment, the most common were antihistamines for both mild (73%) and CS itch (66%), followed by bile acid binding resins (23 and 25%, respectively) (Table 4). Patients with CS itch, as opposed to those with mild itch, were also more likely to have the following concomitant medications: lactulose (16%), spironolactone (15%), pantoprazole (14%), and ondansetron (12%). When examining treatment strategies at sites, patients treated at community sites received UDCA slightly more than those at academic sites (81%, n = 29; 71%, n = 124, respectively) and slightly less combinations of UDCA/OCA (17%, n = 6; 20%, n = 35, respectively) and UDCA/fenofibrate (3%, n = 1; 5%, n = 9, respectively).

Table 4 Overall Treatment by Presence of Pruritus and by Pruritus Severity (on PBC 40 Itch Domain)
Full size table
Fig. 4
figure 4

Current PBC treatments

Full size image

Discussion

These data from TARGET-PBC, a large, real-world US cohort, highlight the pervasive impact of pruritus in patients with PBC, as well as some shortcomings of the medical community’s current response.

The overall prevalence of itch was high, with 81% of respondents reporting itch of any degree, and 37% reporting itch that was severe enough to be deemed “clinically significant” (CS) per PBC-40 scoring. Patients with CS itch were more likely to have advanced disease.

Simultaneous administration of 3 different PROs allowed for the assessment of both congruity and the measurement of the impact of PBC itch on quality of life. Notably, patients with CS itch scored significantly worse on all quality-of-life assessments when compared to those with mild or no itch. Clinically significant itch was associated with worse cognition, fatigue, emotional health, sleep, social life (including isolation, guilt, neglect, and sex life). The impact of itch on quality of life was truly pervasive, and over half had significant fatigue, cognitive, and other general symptom burden.

The association between itch and fatigue among patients with PBC has been previously examined. A study in 2010 focusing on fatigue in PBC showed that 66 patients (20%) indicated pruritus at the time of the administered questionnaire and this was associated with higher fatigue scores than those who did not report itch (32.9 + 11.1 versus 26.0 + 10.8, p < 0.001) [11]. Itch and fatigue both directly impact a patient’s overall quality of life, and it is extremely likely that itch can negatively influence the amount of fatigue a patient reports. Persistent pruritus has been found to impede sleep and lead to severe sleep deprivation [12].

This study found that pruritus in PBC is under-treated in clinical practice. Only half (50.8%) of patients with clinically significant itch were receiving treatment at the time of the surveys, and a third reported never receiving any medical treatment for itch. When itch medication was used, the step-wise guidelines put forth by specialty professional societies was not usually followed. In this cohort, 69.4% of patients with itch were currently treated with antihistamines, despite data that cholestatic itch is not histamine-mediated [13]. Only 24% of those with any itch who were receiving treatment for pruritus, and 9% all patients with any itch reported, were treated with bile acid binding resins, which are recommended as first-line therapy by both AASLD and EASL. Patients with CS itch were more likely to be taking fenofibrate; this may reflect their refractory disease, or the tendency of physicians to prefer fibrates in patients with itch since fibrates have been associated with improvement in itch [14,15,16]. However, it is perhaps surprising that despite treatment with fibrates these patients were still reporting CS itch.

A distinct advantage of these data is that they are derived from a broad, real-world collection of information, both retrospective and prospective. Clinical trial participants were excluded, and the 38 sites were diverse, including both community and academic centers across the USA. Ethnic diversity, while still limited, was increased compared to other PBC studies [17]. Although some selection bias may have occurred because not all enrolled subjects completed the PRO surveys, the percentage of respondents (31%) was good compared to most online survey response rates [18]. Whites were more than twice as likely to complete the PROs as compared to Blacks, Hispanics, and Asians. Survey respondents were also less likely to have advanced disease. By design, PROs were examined cross-sectionally based on availability of data and the treatment efforts were examined throughout the retrospective and prospective period in TARGET-PBC. However, there is a paucity of data examining PROs among patients with PBC in combination with comprehensive, robust data from medical records. These findings help provide a much needed examination into a patient’s quality of life and how it relates to current and past treatment patterns.

Given that clinically significant itch was more often seen in patients with advanced disease, the potential selection bias of this study may have led to an underestimation of the true prevalence and impact of pruritus in the real world. The current study is also not able to elucidate the pathophysiology of cholestatic itching or prove a mechanistic cause and effect between itching and quality of life, but it clearly shows a disease severity-dependent, association. Additionally, OCA, as in the US label, might induce itch which could act as a potential confounder for reported pruritus among the subset of patients currently who were receiving OCA [19]. The study design presented here does not allow for investigating the proportion of patients whose itching started or worsened following the start of OCA and therefore should be investigated further. While TARGET RWE does have information regarding the dose of medication and frequency when it is available within the medical record, information was not obtained regarding patient compliance and the frequency of refills.

These eye-opening data illustrate the real-world extent, impact, and current management practice patterns of pruritus in PBC patients within the USA. Unfortunately, debilitating itch is prevalent but underappreciated, and current options for medical treatment are not fully utilized. We speculate this may be because current treatment options are only partially effective, poorly tolerated, and none are FDA-approved for treatment of cholestatic pruritus in patients with PBC specifically. Rigorously designed clinical trials, as well as greater research efforts, are needed to better evaluate and communicate the debilitating impact of pruritus in PBC patients and to identify highly effective therapies needed to provide effective solutions to this significant problem.

Key Points

For patients enrolled in TARGET-PBC, itching was examined to assess how a patient’s quality of life was impacted. Patients who reported worse itch also reported worse quality of life and more likely to be fatigued as measured by patient-reported surveys. Despite itching being a common problem among patients with PBC, it is not consistently treated with medications.