CORR Insights®: Is There a Cardiotoxicity Associated With Metallic Head Hip Prostheses? A Cohort Study in the French National Health Insurance Databases

Where Are We Now?

In a nationwide retrospective study of 255,350 primary THA patients in France, Lassalle and colleagues [9] identified an excess risk of cardiomyopathy and heart failure in patients with metallic femoral heads compared with ceramic heads, particularly in women and older patients. Importantly, the risk was not just confined to patients with metal-on-metal (MoM) implants, but also included those with metal-on-polyethylene (MoP) implants. So far, this is the fourth observational study of which I am aware that addresses cardiotoxicity with metallic implants [4, 5, 13], and the second to report a positive association [4]. This study trumps the earlier studies in terms of sample size and methodological rigor in addressing confounding, selection and measurement bias, and provides a new impetus to the debate on cardiotoxicity in THA.

Emerging evidence on cardiotoxicity in THA primarily comes from anecdotal case reports [1, 3] with supportive findings from toxicology studies, as well as occupational and environmental exposure studies [2]. Echocardiographic studies are few, and so far, only one cross-sectional study of 70 THA patients reported 7% lower cardiac ejection fraction in THA patients who had metal-on-metal hip resurfacing implants versus patients with MoP implants [11]. Additionally, a postmortem study of 75 patients who had received MoP THAs and matched non-THA controls reported considerably higher myocardial cobalt levels in patients with MoP THAs, particularly in those who had undergone revision procedures [14]. Pathologic changes consistent with cobalt-induced myocardial damage were also more common in THA patients including lower premortem ejection fraction, cardiomegaly, interstitial fibrosis and increased left ventricular wall thickness. Despite the conflicting findings of the four large observational studies, the current body of evidence raises valid concerns about the systemic toxicity of THA implants.

Where Do We Need To Go?

The issue of cardiotoxicity in THA is far from being resolved. Moving forward, there is a need for methodically rigorous observational studies minimizing the effects of confounding and bias. In that respect, the current study is noteworthy for its strict definitions of implant exposure groups in an effort to reduce indication bias, adjustment for a long list of confounders, and sensitivity analyses to examine the robustness of findings. Moreover, observational evidence should be evaluated collectively, rather than one study at a time. For example, indirect evidence from long-term mortality studies is quite relevant [6, 10]. Because healthier individuals often are selected for surgery, mortality is typically lower than the general population within the first 10 years following THA. Yet mortality worsens and exceeds the general population mortality beyond 10 years [10]. About half of the deaths beyond 10 years are cardiovascular deaths, and there is evidence to suggest that the excess long-term mortality is in part due to cardiovascular mortality [6]. In the absence of alternative explanations, it is perhaps reasonable to speculate that the excess cardiovascular mortality is driven, at least in part, by deaths due to cardiomyopathy and heart failure. Although the case reports of heart failure primarily included THA patients with excessively high levels of metal ions, the effects of long-term, low-level chronic exposure are unknown, particularly in vulnerable subgroups, such as older patients with coexisting conditions. By reporting an excess risk of heart failure both with MoM and MoP implants, the study by Lassalle and colleagues [9] underscores the need to study a wide spectrum of implants, and not just those with a presumably higher likelihood of metal exposure.

One fundamental gap in our understanding of potential cardiotoxicity in THA is the potential biological mechanisms. For example, one mechanism is related to the calcium handling because cobalt uses the same transcellular pumps as calcium [7]. Cardiac tissue has the highest calcium influx in the body in conjunction with low cell turnover rates, increasing toxic susceptibility. Although calcium and cobalt are co-transported into cells together, there is not a reciprocal efflux mechanism for cobalt, leading to intracellular accumulation and subsequent damage. Although there are several hypothetical mechanisms, evidence is limited for each.

How Do We Get There?

Observational studies are crucial when examining cardiotoxicity in THA. Yet, they also present several methodological barriers. First, limited infrastructure for long-term safety surveillance in large cohorts of THA patients with continuous long-term followup has traditionally been a major challenge. Fortunately, the landscape is changing on this point, as evidenced by record-linkage studies from Sweden [6], France [9], United Kingdom [13] and Australia [4]. Although cardiotoxicity studies to date have focused on THA cohorts stratified by bearing surface, additional studies are needed, such as the Swedish cardiovascular mortality study [6], comparing cardiovascular outcomes in strictly matched THA and non-THA cohorts. Furthermore, with the ever-increasing usage of THA [8], it is feasible to tap into existing large-scale longitudinal cardiovascular epidemiology studies [12] in order to examine the clinical and hemodynamic characteristics and natural history of heart failure in subgroups of THA patients.

Second, the long latency period between THA surgery and the development of clinical signs of heart failure can obscure the potential contribution of THA. Cardiomyopathy and worsening cardiac function may manifest several years after THA, and establishing causality is challenging with the long induction period between THA surgery and heart failure [1, 3]. Current cardiotoxicity studies [4, 5, 9, 13], have had only about 5-7 years of followup, which may be too short to identify systemic toxicity associated with lifelong chronic exposure. Therefore, robust epidemiological infrastructure with decades-long medical-history data is needed to accurately measure THA exposure and cardiac outcomes that occur several years later.

Third, studies evaluating hospitalizations related to heart failure may underestimate the contribution of THA to the problem because our inability to validate heart-failure-related outcomes potentially leads to non-differential misclassification and biases the risk estimates towards the null. This may be in part why the findings of observational studies are inconsistent. Current THA studies define heart failure outcomes using hospitalizations. Yet, heart failure is increasingly managed in the outpatient setting, and most hospitalizations in patients with are due to noncardiovascular causes creating concerns for outcome misclassification [12]. It is therefore important to pursue research in collaboration with cardiology colleagues who can assist with standardized classification of heart failure outcomes, and interpretation of the etiology and subtypes of heart failure in THA patients.

Fourth, there is a need for adjustment for important confounders such as the presence of obesity, coronary heart disease, hypertension, atrial fibrillation and diabetes.

Fifth, prospective echocardiographic studies with measures of cardiac biomarkers, synovial and systemic metal levels and better documentation of functional and structural changes are important for etiological clues. For example, ejection fraction is preserved in about half of the patients with heart failure, and the etiology and natural history are different than heart failure among patients who present with a reduced ejection fraction [12].

Finally, heart failure is a relatively common condition in older patients. Therefore, prospective longitudinal followup studies with serial measures and predefined cardiac endpoints are feasible. While this would require considerable time commitment and financial support, it would provide a definitive answer to the question of whether THA patients or certain subgroups are at an increased of systemic toxicity.

All told, the jury is still out as to the presence, magnitude, and potential mechanisms of cardiotoxicity in THA. A better understanding of the systemic toxicity of THA implants is important to millions who currently live with THA implants, and those who will choose to undergo THA in the future.