In patients with cyanosis from congenital heart disease, erythropoiesis is governed by many factors that can alter the expected relation between the oxygen saturation (O 2sat ) and hemoglobin concentration. We sought to define the relation between the O 2sat and hemoglobin in such patients and to predict an ideal hemoglobin concentration for a given O 2sat . Adults with congenital heart defects and cyanosis were studied prospectively using blood tests and exercise testing. Nonoptimal hemoglobin was defined as any evidence of inadequate erythropoiesis (i.e., iron, folate, or vitamin B 12 deficiency, increased erythropoietin, reticulocytosis, or a right-shifted oxygen-hemoglobin curve). For patients without these factors, a linear regression equation of hemoglobin versus O 2sat was used to predict the optimal hemoglobin for all patients. Of the 65 patients studied, 21 met all the prestudy criteria for an optimal hemoglobin. For all patients, no correlation was found between O 2sat and hemoglobin (r = −0.22). However, a strong linear correlation was found for those meeting the criteria for optimal hemoglobin (r = −0.865, p <0.001). The optimal hemoglobin regression equation was as follows: predicted hemoglobin = 57.5 − (0.444 × O 2sat ). A negative correlation was found between the hemoglobin difference (predicted minus measured) and exercise duration on cardiopulmonary exercise testing (r = −0.396, p = 0.005) and the 6-minute walk distance (r = −0.468, p <0.001). In conclusion, a strong relation between O 2sat and hemoglobin concentration can be shown in stable cyanotic patients and used to predict an optimal hemoglobin. This relation might be useful in defining functional anemia in this group.
In the clinical care of cyanotic patients with congenital heart disease, it is necessary to assess the appropriateness of a measured hemoglobin level for a given oxygen saturation (O 2sat ). Although multiple factors can influence both hemoglobin and O 2sat , a tool to predict the optimal relation between these variables would be valuable, particularly for situations in which the hemoglobin might be significantly less than expected, such as postoperatively or after severe hemoptysis. We hypothesized that by controlling for factors that might alter this relation, particularly those that might limit erythropoiesis, such as iron deficiency, an ideal linear relation could be found that would define an “optimal” hemoglobin level for a given O 2sat . We also hypothesized that patients with an optimal hemoglobin might have a clinical advantage, as measured by the exertional capacity. We therefore prospectively measured the variables that could potentially alter the hemoglobin–O 2sat association to determine their optimal relation.
Methods
We prospectively enrolled consecutive adults with congenital heart disease in a descriptive cross-sectional study. Patients gave consent, and institutional ethics review approved the protocol. Patients were included if they had a known congenital defect with a right-to-left shunt. We included patients with a wide range of O 2sat , including some patients who had undergone previous repair and had a normal O 2sat at the study. All tests were obtained within a 24-hour period. Other data from the present study have been previously reported.
The patients were recruited and seen at the Royal Brompton Hospital. Additional blood testing was done at King’s College Hospital and St. George’s Hospital (London, United Kingdom). The analyses were performed at the Oregon Health and Sciences University (Portland, Oregon).
O 2sat was measured using transcutaneous spectrometry in the finger after 5 minutes of rest in the sitting position. The blood was drawn in the morning with the patient in a nonfasting state using a venous cannula in the antecubital region. The hemoglobin concentration, packed cell volume, platelet count, basic serum chemistry panels, liver function tests, iron, ferritin, transferrin saturation, red blood cell vitamin B 12 , folate, thyroid stimulating hormone, and serum erythropoietin were measured. The percentage of hypochromic cells and reticulocyte count were measured using an automated Coulter counter (Advia 120, Bayer, United Kingdom). The partial oxygen pressure at half saturation (P50) of the O 2 –hemoglobin dissociation curve was also measured (Hem-O-Scan, American Instrument Company, Silver Springs, Maryland). The whole blood viscosity over a range of shear was measured using a rotational viscometer. The viscosity was then remeasured after the hematocrit had been diluted to 45% using autologous serum. The patients also performed a 6-minute walk test and treadmill exercise, with the measured oxygen consumption and ventilatory efficiency recorded, as previously described.
After the collection of all data, we identified those patients with any evidence of potentially inadequate or excessive erythropoiesis, according to the presence of ≥1 of the following a priori criteria: evidence of iron deficiency, vitamin B 12 or folate deficiency, elevated serum erythropoietin, reticulocytosis, hypochromia, or a significant rightward shift of the O 2 –hemoglobin dissociation curve ( Table 1 ). We also excluded patients using various clinical criteria, including acute hospitalization, therapeutic phlebotomy within the previous 6 months, and recent significant hemoptysis (requiring hospitalization). Patients with a patent ductus arteriosus and differential cyanosis were also excluded from the optimal category because of the uncertainty of what the mean O 2sat would be. Patients using supplemental oxygen regularly were excluded because their O 2sat at room air might not have accurately reflected their average daily saturation.
| Variable | Cutoff | Normal Range | Excluded Patients (n) |
|---|---|---|---|
| Transferrin saturation | <20% | 20–45 | 26 |
| Red blood cell folate (μg/L) | <200 | 164–900 | 0 |
| Vitamin B 12 (ng/L) | <180 | 180–900 | 2 |
| Serum erythropoietin (IU/L) | >25 | variable | 9 |
| Reticulocyte count (%) | >2 | <2% | 8 |
| Hypochromic cell count (%) | >6 | <6% | 7 |
| P50 of oxygen–hemoglobin dissociation curve (mm Hg) | >29 | 25–29 | 10 |
After exclusion of any patient who had met these criteria, a plot of the O 2sat and measured hemoglobin was made. A linear regression equation was defined, together with confidence intervals, around this line. Using the regression equation, the values for the predicted hemoglobin were made, and the difference between the predicted and measured hemoglobin was obtained (hemoglobin difference) for each patient. The clinical variables between the patients with and without an optimal hemoglobin level were compared using Student’s t test and correlations using Pearson’s coefficient. The data are expressed as the mean ± SD, and p <0.05 was considered statistically significant. No adjustment was made for multiple comparisons.
Results
A total of 65 patients were studied (mean age 36 ± 12 years, 67% women). For the whole group, the O 2sat at rest was 81 ± 8%, hemoglobin was 19.6 ± 2.9 g/dl, and hematocrit was 60 ± 8%. The anatomic diagnoses are listed in Table 2 , as well as their group designation. All but 1 patient had pulmonary vascular disease. Of the 65 patients, 20 had likely been cyanotic at birth, with 45 (largely with simple shunts) having developed right to left shunting over time. No patient was found to have significant renal, liver, or thyroid dysfunction.
| Diagnosis | Patients (n) | Cyanotic Since Birth | Cyanosis Developed ⁎ | Excluded (Nonoptimal) | Included (Optimal) |
|---|---|---|---|---|---|
| Atrial septal defect | 4 | 0 | 4 | 2 | 2 |
| Ventricular septal defect | 27 | 2 | 25 | 11 | 16 |
| Atrioventricular septal defect | 8 | 0 | 8 | 7 | 1 |
| Patent ductus arteriosus | 8 | 0 | 8 | 8 | 0 |
| Truncus arteriosus | 6 | 6 | 0 | 6 | 0 |
| Transposition of great arteries | 5 | 5 | 0 | 4 | 1 |
| Pulmonary atresia with ventricular septal defect | 2 | 2 | 0 | 2 | 0 |
| Other complex | 5 | 5 | 0 | 4 | 1 |
| Total | 65 | 20 | 45 | 44 | 21 |
⁎ Developed right-to-left shunt in setting of elevated pulmonary vascular resistance (Eisenmenger reaction).
Of the 65 patients studied, 44 met ≥1 criteria for exclusion, and most of them met at least 2 criteria ( Table 1 ). The exclusions on clinical grounds included 16 using supplemental oxygen, 8 with differential cyanosis, 10 with recent phlebotomy, and 2 with recent hemoptysis. The most common exclusion criterion was iron deficiency, and many of those patients met additional criteria associated with iron deficiency (e.g., phlebotomy, hemoptysis, P50 shift, or erythropoietin elevation). After the exclusions, 21 patients had met all the criteria for adequate erythropoiesis.
For the entire cohort, no significant relation was found between O 2sat and hemoglobin ( Figure 1 ). In contrast, when patients with evidence of inadequate erythropoiesis were excluded, a strong linear relation was found. The slope and intercept for the regression line defined a predicted optimal hemoglobin as follows: predicted hemoglobin = –0.444(O 2sat ) + 57.5.
From this, the predicted hemoglobin for a given O 2sat , including the upper and lower confidence intervals, were calculated ( Table 3 and Figure 2 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
