Hemoglobin Decline and Occult Cancer After OAC Initiation

Corresponding Author

Oral anticoagulation (OAC) is the cornerstone of medical therapy for the prevention of stroke and systemic thromboembolism in patients with atrial fibrillation (AF).¹ In clinical practice, a decline in hemoglobin (Hb) levels without overt bleeding is not uncommon among patients with AF receiving OAC. Such changes are often attributed to aging, frailty, comorbidities, laboratory variability, or anticoagulation-related blood loss and may consequently be overlooked. However, a decline in Hb may be an early, clinically meaningful signal of occult bleeding and portend future clinically relevant bleeding events.

Previous large population-based studies have suggested that overt bleeding after OAC initiation may be the first sign of undiagnosed cancer, particularly when bleeding originates from the gastrointestinal (GI), genitourinary, or respiratory tract.2, 3, 4 In a recent cohort of approximately 120,000 patients with AF receiving OAC, bleeding was associated with a 4-fold higher risk of a new cancer diagnosis and with cancer detection at an earlier stage.² These findings suggest that OAC therapy may lower the threshold for bleeding from an occult lesion, thereby acting as a “bleeding stress test.” Accordingly, vigilant evaluation for occult cancer may be warranted in patients with overt bleeding. However, it remains unclear whether a gradual or unexplained decline in hemoglobin, even in the absence of clinically relevant bleeding, is associated with an increased risk of undiagnosed cancer.

In this issue of JACC: CardioOncology, Yang et al⁵ address this clinically relevant question. Using the Korean National Health Insurance Service database linked to serial national health examinations, the investigators evaluated 6,789 patients with AF who newly initiated OAC and underwent health examinations with Hb measurements within 1 year before and after OAC initiation. They assessed the association between changes in Hb levels and the subsequent risk of newly diagnosed cancer. A significant Hb decline (≥2 g/dL) was observed in 7.8% of patients and was associated with a higher incidence of newly diagnosed cancer (6.7 vs 4.3 per 100 person-years; adjusted HR: 1.42; 95% CI: 1.10-1.97). This association was more pronounced for GI cancer (adjusted HR: 1.78; 95% CI: 1.12-2.83). The key findings remained consistent across competing-risk, propensity score–weighted, and several sensitivity analyses. Patients with an Hb decline also had higher risks of all-cause mortality and major bleeding, but not of stroke or systemic embolism.

The authors should be commended for their extensive analytic work. This study provides 2 clinically relevant insights at the intersection of OAC therapy, Hb decline, and cancer detection. First, it shifts the detection window upstream—from overt bleeding to a simple laboratory signal that may otherwise be overlooked. A gradual decline in Hb may serve as a sentinel sign of an occult, bleeding-prone malignancy before clinically apparent bleeding occurs. Second, the findings support the concept of OAC as a “bleeding stress test” that may detect more vulnerable patients and unmask occult cancer. The graded association with declining Hb, the excess risk of GI cancer, and the short interval to cancer diagnosis collectively support the unmasking of preexisting malignancy rather than de novo cancer development caused by OAC—an interpretation consistent with the absence of an established causal link between OAC and cancer.⁶

However, several methodological limitations and interpretive caveats should be acknowledged. First, this study was based on national administrative claims data, which are inherently susceptible to coding errors, missing information, and a lack of clinically relevant details, including coexisting clinical variables and concomitant use of over-the-counter medications that may not be captured in such data sources. The study also may be subject to misclassification and ascertainment biases. For example, the unexpectedly high prevalence of several comorbidities, including heart failure (57%), hypertension (89%), liver disease (46%), and previous stroke or transient ischemic attack (38%), raises concerns regarding coding-based comorbidity ascertainment and the representativeness of the study population relative to contemporary prospective AF registries with systematically collected clinical data.^6,7 Furthermore, despite robust propensity-score methods and multiple sensitivity analyses, residual confounding from unmeasured factors cannot be excluded. Second, cancer diagnoses and clinical events were not centrally adjudicated, leaving the possibility of endpoint misclassification. Third, because national health examinations include cancer-screening procedures, such as fecal occult blood testing and endoscopy, an observed decline in Hb may have prompted additional diagnostic evaluation, thereby increasing the likelihood of cancer detection. This potential surveillance bias may have contributed to the observed association. Fourth, the median follow-up duration for cancer detection was relatively short (10.6 months). Although the short interval may support the unmasking of preexisting occult malignancy, it remains uncertain whether Hb decline is a reliable marker of cancer risk over longer-term follow-up. Fifth, the study lacked comparator groups without AF or without OAC therapy. Therefore, it remains unclear whether the association between Hb decline and subsequent cancer detection is specific to patients with AF receiving OAC therapy or represents a more general association that also applies to patients with other medical conditions or even to the general healthy population. Finally, the study was conducted in a single country. Therefore, the applicability and generalizability of these findings to populations with different ethnic, health care, and clinical backgrounds remain uncertain.

What are the clinical implications of the current study? In daily practice, a decline in Hb should prompt assessment for bleeding symptoms, iron deficiency, chronic kidney disease, inflammatory conditions, and concomitant use of antiplatelet agents or nonsteroidal anti-inflammatory drugs. Further evaluation for occult GI blood loss, including endoscopic assessment when clinically appropriate, may be considered when the Hb decline is substantial, progressive, or otherwise unexplained, particularly in the presence of iron deficiency or related GI symptoms.^4,8 In addition, additional large-sized prospective studies should determine which Hb trajectories warrant further investigation and whether Hb-guided diagnostic strategies improve cancer detection and clinical outcomes while minimizing unnecessary diagnostic testing. For now, Yang et al⁵ provide an important clinical reminder: in anticoagulated patients with AF, a decline in Hb—even in the absence of overt bleeding—should not be overlooked. Although not every decline signals occult cancer, there may indeed be “no smoke without fire.”

Funding Support and Author Disclosures

Dr Park has received research grants and speaker fees from Abbott Vascular, Boston Scientific, Daiichi-Sankyo, Yuhan Pharm, Daewoong Pharm, Edwards Lifesciences, and ChongKunDang Pharm. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.