New challenges for an expanding generation of older persons with haemophilia

Data for this study were sourced from CHESS, a cross-sectional retrospective study of patients with severe haemophilia A or B in France, Germany, Italy, Spain and the United Kingdom. Data were collected through two questionnaires. The first was administered to physicians, who were asked to provide information on direct medical resource utilisation and clinical data for their haemophilia patients via a web-based case record form (CRF). The physicians asked the same patients to complete corresponding patient self-completion (PSC) questionnaires covering information on patient reported outcomes that determine HRQoL. Data were collected between December 2014 and April 2015.

Patients with a current inhibitor diagnosis were excluded (n=58) but those with a previous diagnosis were included and followed up as per the rest of the study population. The study sample included CRF data for 1,227 patients, of whom 523 (43%) completed the corresponding PSC. In CHESS, factor consumption was documented by the physician in the CRF. The treatment strategies were categorised as: prophylaxis from diagnosis (PX), on prophylaxis, previously on-demand (PXOD), always been on-demand (AOD) and previously on prophylaxis and moved to on-demand regimen (PPOD). For on-demand regimens, factor consumption for the most recent three-month period was annualised; for prophylaxis regimens, mean IU per infusion was multiplied by the weekly infusion rate, and annualised.

Bleeding frequency reported in this study was a combination of minor and major bleeding events occurring in the preceding 12-month period prior to enrolment in the study and recorded in the CRF (annual bleeding rate, ABR). A minor bleed was defined as: mild pain, minimal swelling, minimal restrictions of motion and resolution within 24 hours of treatment. A major bleed was defined as pain, effusion, limitation of motion and failure to respond within 24 hours.

The International Society on Thrombosis and Haemostasis defines a target joint as three or more spontaneous bleeds into a single joint within six consecutive months; joints meeting that definition which subsequently have two or fewer bleeds in 12 consecutive months are no longer considered target joints ^[1]. In CHESS, a target joint was defined by the clinicians, taking into account bleeding frequency and period of observation; this approach means that joints affected by bleeding are included when they no longer meeting the criteria for a target joint but have been affected by chronic synovitis secondary to prior joint bleeding. For clarity, joints identified in this way will be described as ‘affected joints’. The number of affected joints for each patient was recorded in the CRF.

We recorded the presence of comorbidities listed in the medical records of each patient, using categories defined in the CRF. HRQoL was measured using the EQ-5D, a self-administered, generic, preference-based instrument designed to measure the impact of disease on an individual's health state. The instrument comprises the EQ-5D health state descriptive; and the EQ Visual Analogue Scale (EQ-VAS) ^[10]. The EQ-5D instrument consists of five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has three levels: no problems, some problems, and extreme problems. The responses across these five dimensions provide a description of the patient's current health state. As is standard practice, we compared EQ-5D values for the multinational CHESS population with those of a single state, in this case the UK, which had the biggest cohort in the CHESS population. Preference-based values for the EQ-5D health states were obtained using values associated with the health states derived in a UK population (tariff values) ^[11]. These UK-specific tariff values were applied to all patients in our sample, regardless of their country of origin. This approach ensures that additional variation due to differences in tariff values between countries was not introduced into the analysis. This allows us to compare each patient using the same tariff value, so that changes in EQ-5D are determined only by changes in domain scores. The EQ-VAS (visual analogue scale) is measured on a vertical visual analogue scale where the endpoints are described as ‘best imaginable health state’ and ‘worst imaginable health state’.

We used a comparative general population sample from the 2012 Health Survey for England (HSE) ^[9]. This is an annual survey examining changes in the health and lifestyles of the population that includes EQ-5D domain scores and EQ-VAS scores. From this survey population, we selected a sub-sample mirroring CHESS of males aged over 18 years old. The UK population tariff values that were applied to the CHESS data were also applied to the HSE domain scores.

The unit of analysis was patient-level. Demographic and clinical data are presented by age groups (18–30, 31–40, 41–50, 51–60 and 61+). Means, medians and standard deviations or standard errors, where appropriate, were used to describe continuous variables; categorical variables are described as frequencies and proportions. Kruskal-Wallis and Chi-squared statistical tests were conducted in order to test for between-age-group differences based on the type of variable included. All analysis was performed in R version 3.3.3 ^[12].

All patients provided informed consent. The study protocol was approved by the Research Ethics Sub Committee of the Faculty of Health and Social Care within the University of Chester, UK.

Key differences between age groups were observed in the reported current treatment strategy in our study, where only patients in the 18–30 age group were categorised as having prophylaxis from diagnosis (PX) at the study recruitment date (n=217). This is reflective of what was reported by participating physicians; however, given the relatively novel nature of primary prophylaxis at the time, it is likely this is made up of patients who initiated prophylaxis at an early age (a mix of primary and other prophylaxis).

The mean ABR associated with each treatment type was significantly different only in the age groups 18–30 and 61+ (Kruskal-Wallis test, p-value <0.000 and 0.023 respectively). Table 2 presents the mean and median number of minor and major bleeds and ABR reported in the previous 12 months according to treatment strategy by age category. Overall, patients currently receiving on demand therapy (AOD, PPOD) had higher ABR than those receiving any prophylaxis therapy (PX, PXOD). Median ABR associated with PXOD and PPOD for the age groups of 18–30 years and 61+ years were the same. Median ABR associated with PXOD was higher than with PPOD for the age group 41–50 and 51–60 years. PXOD was consistently associated with higher ABR than PX, with rates similar to those associated with current on-demand treatment.

Figure 1 represents a box plot of the medians and ranges of the number of affected joints by age group. This shows that the medians were the same for each age group (1 affected joint) with the exception of age group 61+ (median 0). The range of number of affected joints increases with age. Younger age groups (18–30 and 31–40) had on average 1.00 ± 0.056 and 1.17 ± 0.079 affected joints. By contrast, patients aged 41–50 and 51–60 years had the highest mean number of affected joints at 1.25 ± 0.105 and 1.41 ± 0.174, respectively. As occurred with number of total bleeds, the mean number of affected joints was lower in the 61+ age group (0.9 ± 0.116) and was significantly lower than in the 51–60 age group (p=0.037).

Figure 1

Figure 2

Table 3 lists the prevalence of each of the comorbid conditions. The prevalence of comorbidity increased with age. While a significant proportion of our sample recorded no comorbidities (n=624), the prevalence of comorbidity increased from 36% in the 18–30 age group to 68% in the 61+ age group. The mean number of comorbidities for each age group ranged from 0.55 ± 0.037 for the 18–30 age group to 1.50 ± 0.176 in the over-40 group. Figure 2 shows that the median number of comorbidities increases significantly with age group up to age 60.

The three most frequently reported conditions were anxiety, depression and hypertension. The prevalence of multimorbidity, defined as the presence of two or more chronic medical conditions in addition to haemophilia, was most frequent in patients with osteoporosis (n=16) and rheumatoid arthritis (n=12), in whom the mean number of additional conditions was three or more. Anxiety and depression were the most prevalent morbidities occurring together (n=36) followed by hypertension and hypercholesterolaemia (n=10).

Table 4 presents the proportion of both the CHESS and HSE sample within each of the levels (no problems, some problems, unable/extreme problems) for the EQ-5D domains (mobility, self-care, unusual activities, pain and anxiety/depression). Overall, a pattern of increasing impact on quality of life is noted in line with age for both the CHESS and HSE samples.

The proportion of patients in the CHESS population who reported some problems in the mobility domain increased with age; a similar trend was evident with extreme problems but comparatively few people were affected. For example, 63.5% of individuals in the 61+ age group reported some problems but 21.9% of those in the 18–30 age group did so. A similar pattern is apparent in the self-care domain: the proportion reporting extreme problems was consistently low across age-groups but the proportion of CHESS patients reporting some problems increased at 30–39 years (20.8%) and more than doubled among those aged 61+ (44.2%). Anxiety or depression affected 30% of 18–30-year-olds and over 60% of the 61+ group, with extreme problems reported by one in six of the older age group who were affected. The proportion reporting some or extreme problems in the pain domain was greater than for any other domain and high across all age groups. Some problems were reported by 45.7% in the 18–30 age group (vs. 7.3% for self-care and 28.1% for anxiety/depression in the same age group). Almost 50% of people in the age group 18–30 reported some or extreme problems in the pain domain.

In the HSE population, 31.3% of patients in the 61+ age group reported some mobility problems whereas only 4.3% of the 18–30 age group did so. A small proportion of people in the HSE population reported problems with self-care and this increased only slightly with age. Problems with anxiety or depression were reported by 13.9% of the 18–30 age group and peaked at 20.7% among 51–60-year-olds. Pain was again the most frequently reported issue and prevalence increased with age, reaching 41.3% with some problems and 5.5% with extreme problems in the 61+ age group.

Both the EQ-5D index score and the EQ-VAS are illustrated in Figure 3 for both the CHESS and HSE sample populations. A consistent downward trend is noted for the EQ-5D index across age groups in both populations, with marked inter-individual variation. In CHESS, the mean index ranged from 0.82 in the 18–30 age group to 0.57 in the 61+ group (Table 4). The EQ-VAS, which is a self-reported measure, largely followed the same patterns. In the CHESS population, mean EQ-VAS score was 75.21 in the 18–30 age group and 60.92 in the 61+ group. In the HSE population, the mean EQ-5D index ranged from 0.94 among 18–30-year-olds and 0.81 in the 61+ group; the mean EQ-VAS among 18–30-year-olds was 81.98 and in the 61+ group it was 74.30.

Figure 3