Anticoagulation management and poor clinical outcomes in tamariki and rangatahi with rheumatic heart disease following mechanical valve replacement surgery in Counties Manukau

Rheumatic heart disease (RHD) is the chronic sequela of acute rheumatic fever (ARF), a multisystem inflammatory condition that develops in a susceptible host due to Group A Streptococcus pharyngeal infection.¹ In 2019, the global burden of RHD was estimated to be 40.5 million cases, mostly in low- and middle-income countries.^2,3 Despite Aotearoa New Zealand being a developed country, RHD is a significant cause of cardiovascular morbidity and premature mortality among Māori and Pacific peoples.^4,5 Serious consequences of RHD include infective endocarditis, atrial fibrillation, thromboembolism, stroke and premature death.⁶

In Aotearoa, RHD prevalence is estimated to be around 2% among Pacific tamariki and rangatahi, and 1% among Māori.^7–10 The New Zealand Rheumatic Heart Disease Registry Project estimates there are around 3,600 individuals with moderate or severe RHD.⁵ ARF and RHD hospitalisation costs in Aotearoa peak between the ages of 5 and 19 years.¹¹ The most recent analysis of direct hospitalisation costs for ARF/RHD in young people was published in 2012 and found an average annual hospitalisation cost of NZ$12 million, with heart valve surgery accounting for 71% of hospitalisation costs for those between 5 and 14 years of age.¹² Around 160 deaths in Aotearoa annually can be attributed to RHD.¹²

A proportion of tamariki and rangatahi with severe RHD require heart valve surgery. In those <16 years, surgery is performed at Starship Hospital, whereas older patients usually undergo surgery in adult cardiosurgical centres. Where possible, valve repair procedures are favoured over mechanical valve replacement in order to avoid the need for lifelong anticoagulation.^13,14 Valve repair surgery in Aotearoa has excellent outcomes with 90% survival and 75% freedom from reoperation at 10 years.¹⁴ However, in some situations mechanical valve replacement surgery is required due to the unrepairable state of the native valve.

Mechanical valve surgery necessitates lifelong oral anticoagulation therapy with the vitamin K antagonist warfarin to prevent thromboembolic events.¹⁵ Tamariki and rangatahi face significant challenges with long-term warfarin therapy due to the complexity of international normalised ratio (INR) monitoring, the need for frequent warfarin dose adjustments and the effect of diet, medication interactions, illness and missed doses on INR levels.¹⁶ Furthermore, for women in childbearing years, warfarin is teratogenic and can lead to bleeding complications during pregnancy, along with increased risk of major thromboembolic events.^17,18 The challenges are particularly profound for rangatahi during the transition from paediatric to adult health services. In the Auckland Region, patients are typically transitioned from the Starship service to primary care providers around the age of 16 years. CoaguChek® machines and test strips are not routinely available or funded for adult patients in Aotearoa. Currently there are two main methods for primary care INR management: firstly, via general practitioner (GP) and community laboratory, and secondly, via the community pharmacy anticoagulation monitoring service (CPAMS).¹⁹ Further challenges in the journey towards independence and self-management can include frequently changing addresses and busy work and study schedules. There are little published data about anticoagulation-related outcomes in young people with RHD. To date, there have been no Aotearoa studies on this subject.

Aims

The primary objective of this study was to describe the patient characteristics of all tamariki and rangatahi with RHD on anticoagulation after mechanical valve surgery in Counties Manukau and to review anticoagulation management and complications. Secondary objectives included describing ARF recurrences in this population, and describing deaths in those ≤25 years following RHD mechanical valve surgery occurring between January 2016 and January 2021.

Methods

This study was undertaken in Counties Manukau, the district with the highest prevalence of RHD in Aotearoa.

We undertook a retrospective review of all tamariki and rangatahi aged ≤25 years on anticoagulation after metal mechanical valve replacement for RHD, domiciled in Counties Manukau between 1 January 2016 and 1 January 2021. Patients with RHD and on anticoagulation for other reasons, such as heart failure or atrial fibrillation, and patients with valve repairs were excluded. Those with only bioprosthetic valves were excluded. Data census was conducted on 1 January 2021, or at the time of last documentation in clinical records for those who had died.

The following variables were investigated: age as of 31 January 2021, gender, ethnicity and age at RHD diagnosis. Clinical variables included mechanical valve surgery details, anticoagulation monitoring and hospitalisations due to subtherapeutic INR without clinical complication, supratherapeutic INR without clinical complication, haemorrhage, stroke, other thromboembolic events and mechanical valve thrombosis. Cardiology and rheumatic fever clinic attendance, medication dispensing records, transition process, age of transfer to adult services and clinical details of deaths were also examined. Hospitalisation events were defined as events where the patient was admitted to an inpatient service and stayed overnight. Emergency department visits were excluded.

Multiple data sources were interrogated to optimise case ascertainment, including the New Zealand Rheumatic Heart Disease Registry Project,⁵ Counties Manukau rheumatic fever secondary prophylaxis database, Starship Hospital anticoagulation database and clinician records. Electronic records and primary care prescribing records were evaluated.

Continuous variables are presented as median and range. Categorical data are presented as frequencies and percentages. Statistical analyses were performed using Microsoft Excel.

Ethics approval was granted by the Auckland Health Research Ethics Committee (AH22290). Locality approval was obtained from the Counties Manukau Health Research Office.

Results

There were 53 patients in Counties Manukau with a mechanical metal valve on anticoagulation secondary to RHD between 1 January 2016 and 1 January 2021. Review of clinical records identified that five of these individuals died sometime between 1 January 2016 and 1 January 2021.

Patient demographics are presented in Table 1. Of the 48 patients alive in 2021, the majority were aged ≥16 years (42/48, 88%). There was a male predominance (58%). The highest represented ethnicities were Samoan (39%), Cook Island Māori (21%), Māori (19%) and Tongan (17%).

View Table 1–3.

The median age at time of first mechanical valve surgery was 15 years (range 4–23).

A description of the valves in surgical management of affected valves is presented in Table 2. The aortic valve was the most commonly operated on valve (35%), and the sole mitral valve was the least common (10%). The majority of the cohort had one or two valves operated on (44% and 41%, respectively). On-X was the most frequently chosen type of mechanical valve (62%).

There were 43 patients under the care of adult services and 10 patients under the care of paediatric services. The median age of transition from paediatric to adult services was 16 years (range 15–19). Of the 43 patients under adult services, 33 patients (62%) had been transferred to adult services via formal referral from paediatrics, and seven patients (13%) underwent surgery and commenced warfarin under adult services. Five patients in the cohort had become lost to follow-up during or after transition and were re-referred to adult cardiology services either following an acute hospital presentation or by primary care.

The median duration of anticoagulation was 4 years (range 0.5–18 years). The most common mode of anticoagulation monitoring in the past year was via the community laboratory service and GP (28, 53%), followed by CPAMS point-of-care testing (11, 21%), self-testing with CoaguChek® machines and paediatric cardiac nurse specialist input (8, 14%) and monitoring at Middlemore Hospital (5, 10%). One patient had point-of-care testing at a GP practice. Of the five patients who monitored INR at the hospital, two were on renal dialysis, two had prolonged hospitalisations and one infrequently monitored INR intermittently during hospital admissions.

During the 5-year study period, there were 80 anticoagulation-related hospitalisations, occurring in 38 individuals (72% of the cohort).

Causes for hospitalisation included subtherapeutic INR without clinical complication (42,52%), INR without clinical complication (12, 15%), haemorrhage (11, 14%), stroke (7, 9%), other thromboembolic events (5, 6%) and prosthetic valve thrombosis (3, 4%). One pregnant patient experienced foetal demise. Among the 53 individuals, 38 (72%) had ≥1 hospitalisation, and 20 individuals had two or more hospitalisations. Thirty-three percent of hospitalisations were due to significant complications such as haemorrhage, stroke, other thromboembolic events and prosthetic valve thrombosis. Of the 26 hospitalisations for a major clinical complication, 15 individuals (58%) had a prior admission with supratherapeutic or subtherapeutic INR.

Eleven hospitalisations secondary to haemorrhage included three patients with severe menorrhagia, four patients with epistaxis and one patient each with adnexal bleeding, per rectal bleeding and knee haemarthrosis. Other thromboembolic events were limb ischaemia, splenic infarction, renal pole infarction, cardioembolic myocardial infarction and multiple small emboli. All three cases of prosthetic valve thrombosis required prolonged hospital admissions. All the subtherapeutic and supratherapeutic INR admissions without clinical complications required bridging anticoagulation or administering vitamin K with observation.

Treatment adherence was challenging to determine from the available records. Medical dispensing records were used as a surrogate marker. Around half the cohort (21/45 individuals where dispensing records were available) had medication dispensed <5 times in the preceding 12 months.

The majority of patients (37/53 [70%]) presented with ARF, and the median age at ARF diagnosis was 10 years (range 3–21 years). The remaining 16 patients presented with established RHD. Six patients (11%) had a documented ARF recurrence during the 5-year study period. All six cases with recurrent ARF had a history of benzathine penicillin non-adherence. Three ARF recurrences occurred post-mechanical valve replacement.

Five deaths occurred between January 2016 and January 2021: two in the community and three in hospital. The age at death for both community cases was 17 years. The cause of death was reported as “direct complications of RHD”, and warfarin non-adherence was mentioned in clinic letters. One had a sudden collapse, severe aortic regurgitation and long QT interval; the other had an enlarged heart. Of note, post-mortem computed tomography scans of both brains did not reveal intracranial bleeding. Three deaths occurred in hospital. All three patients who died in hospital had prolonged intensive care unit admissions, difficult anticoagulation control and poor cardiac function.

Discussion

To our knowledge, this is the first published report to describe anticoagulation management practices and anticoagulation-related morbidity and mortality in tamariki and rangatahi following mechanical valve surgery for RHD in Aotearoa. Our study highlights the high rates of serious thromboembolic complications such as stroke, bleeding and valve thrombosis and premature deaths in this cohort.

 Among the 53 tamariki and rangatahi in this study, almost three-quarters had at least one hospitalisation. Over half of those who experienced significant complications had a prior hospitalisation for subtherapeutic or supratherapeutic INR, indicating that current services are failing to optimise management and avert future complications. These findings are likely to be a conservative assessment of overall morbidity, as a small number of complications may have been managed in the emergency department without hospitalisation, particularly among older patients in the cohort. There were five RHD-related deaths in rangatahi between 2016 and 2021 in Counties Manukau. These deaths may well have been prevented if accessible, culturally responsive clinical services and enhanced patient support were available.

We found that RHD patients used a variety of anticoagulation services. The most common method of anticoagulation monitoring was via the GP and community laboratory service, followed by CPAMS. While CPAMS has been previously found to provide a high standard of anticoagulation management in the general population on warfarin,¹⁹ only 21% of this RHD cohort utilised CPAMS. This may be due stringent referral criteria, which require the person to have three in-range INR recordings from the GP, prior to referral to the CPAMS service. A small number of younger patients were under the care of the national paediatric and congenital cardiac service (PCCS) at Starship Children’s Hospital. Under PCCS, patients are provided with a free CoaguChek® machine and test strips. They are closely supported by nurse specialists to conduct home-based INR testing, with dosing advice provided by the nurse specialist team. Previous systematic reviews demonstrate that self-monitoring in this way is safe and results in reduced thromboembolic events and more time in the therapeutic range of INR compared to other monitoring modes.^20,21

It is notable that none of the available services are integrated with the Counties Manukau rheumatic fever benzathine penicillin secondary prophylaxis service. In future, flexible referral pathways and widened access to point-of-care INR testing devices for GP clinics and community nurses who deliver benzathine penicillin may improve anticoagulation management options for rangatahi.

 It is well recognised that patient experience of ARF and RHD clinical services in many parts of Aotearoa is poor, with service access and both interpersonal and systemic racism contributing to their poor experience.²² Future anticoagulation services need to be designed taking rangatahi perspectives into consideration, and there is a need for qualitative research and quality of life measures to inform service design. Further work is needed to determine what tamariki and rangatahi would find useful to help support them with anticoagulation management. Culturally responsive frameworks, such as those developed in the recently completed Ministry of Health co-design initiative, should be used when developing anticoagulation services for this group of patients.²³ Tamariki and rangatahi with RHD have numerous contacts with healthcare services and there are opportunities for future models to better integrate with other aspects of primary healthcare and ARF/RHD healthcare.²⁴

The high rate of complications in this cohort also highlights the importance of careful pre-operative planning for RHD valve surgery. Wherever technically possible, valve repair techniques should be used. This is particularly important for young women of childbearing potential and may also be important for young people who play high-level contact sports. Ideally, a multidisciplinary team experience in RHD management, and where possible patient preference, should inform surgical planning.^13,14

This was a single-centre study, and as such findings may not be widely generalisable beyond South Auckland. However, poor outcomes in this cohort highlight the need to understand anticoagulation service models and outcomes around the country for other RHD patients following valve replacement surgery. A national stocktake of ARF/RHF tertiary services would identify regional variation in services, and it is likely that patients living in rural areas and under-served urban centres experience the greatest barriers to accessing anticoagulation services.