Accurate plasma hemoglobin (PHb) measurements are needed to investigate a model proposing a link between intravascular hemolysis and decreased nitric oxide (NO) bioavailability in sickle cell disease (SCD). Artifactual hemolysis hypothetically can produce excessive noise that might obscure the signal and lead to misinterpretation of data. We systematically studied possible blood processing variables that could cause artifactual hemolysis to obscure in vivo levels of free hemoglobin in adults with SCD.
We systematically compared the effects on PHb measured by ELISA of anticoagulant type, Vacutainer materials and processing time on the same blood specimens from healthy controls and adults with SCD divided into aliquots that were then processed using a variety of in vitro conditions. We analyzed the results of PHb measurements from archived specimens from 467 SCD patients. We evaluated in greater detail plasma specimens collected and processed at bedside from 44 patients with SCD under optimal conditions: drawn from indwelling wide-bore brachial artery catheters by minimal negative pressure into heparinized syringes, gently transferred to tubes for centrifugation within 5 minutes, evaluating the associations of PHb levels from these optimally processed samples with relevant in vivo physiological measurements.
Analysis of specimens from healthy control subjects (n=10) and adults with SCD (n=10) for each experiment, split into aliquots to test each variable under defined conditions yielded significant evidence indicating that: (1) arterial catheter syringe specimens are superior to conventional Vacutainer venipuncture specimens; (2) heparin is superior to EDTA anticoagulant; (3) glass tubes are superior to plastic tubes; (4) processing in 2 hours or less is superior to longer processing times. Comparison of 467 archived EDTA plasma specimens processed under comparably variable conditions provided results consistent with the bench top results, confirming the importance of these processing variables on research specimens. Conventional venipuncture EDTA Vacutainer tubes with conventional processing yielded median PHb levels four-fold that of arterial catheter heparin syringe specimens processed at bedside, suggesting that on average 75% of the PHb in conventional specimens may represent background noise and only 25% of the PHb level may represent in vivo free hemoglobin.
Blood specimens were collected and processed by the optimal conditioned defined by the preceding experiments from 44 adults with sickle cell anemia. This set of specimens yielded a median PHb level of 6.07μM (CI95=5.27–7.03μM). PHb correlated significantly with blunting of vasodilation in vivo after brachial artery infusion of each of three graded doses of the NO donor sodium nitroprusside (SNP), an indicator of vascular NO responsiveness; at doses of 0.8μg/min(r=−0.46 p=0.0022), 1.6μg/min (r=−0.41, p=0.0082), and 3.2μg/min (r=−0.46, p=0.0024). Quartile analysis revealed a significant difference in forearm blood flow response across all doses of SNP among the highest, middle, and lowest PHb quartiles (p=0.0386). PHb also correlated significantly with tricuspid regurgitant velocity (R=0.37, p=0.0147), suggesting a significant association with Doppler echocardiography estimates of pulmonary artery pressure. A limited number of patients from this specimen pool (n=7) subsequently underwent clinically indicated right heart catheterization. There was a trend in this small subset correlating PHb with mean pulmonary artery pressure (r=0.61, p=0.1482).
Plasma hemoglobin levels collected and processed under optimal conditions show statistically significant associations with in vivo physiological measures of nitric oxide responsiveness and echocardiographically estimated pulmonary artery pressure, with supportive trends from right heart catheterization in a limited subgroup of patients with SCD. This supports a role for plasma hemoglobin as a biomarker of nitric oxide resistance and pulmonary hypertension in SCD. Conventional phlebotomy into Vacutainer tubes introduces a high background level of artifactual ex vivo hemolysis in SCD specimens that is capable of obscuring these associations, especially if processing is delayed 2 hours or more.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.