Author + information
- John E Madias, MD, FACC* ()
- ↵*Professor of Medicine Chief, Cardiology Division, Elmhurst Hospital Center, 79-01 Broadway, Elmhurst, New York 11373, USA
I am grateful to Dr. Brennan for directing me to literature describing a direct correlation of albumin and amplitude of QRS complexes (1). To respond to his inquiry I have modeled an analysis of our data after the study of Dr. Heaf (1). Our patients were critically ill, were followed for a long time (34.5 ± 36.7 days) (2)and inevitably experienced gradual reduction of albumin, despite our concerted efforts to use enteral and/or parenteral feeding. We had not measured albumin daily, but frequently. Thus, I was able to employ values that were the closest to the admission and peak weight albumin values of our patients. The albumin levels available for my analysis were checked on days 0.61 ± 1.71 for admission, and on days 0.43 ± 1.79 of peak weight time point.
Correlation of the sums of all 12 QRS leads (ΣQRS) (2)from admission with the corresponding albumin values revealed an r = 0.057 and a p = 0.77. A similar analysis, carried out for the ΣQRSs obtained on the day of the peak weight and the corresponding albumin values, showed an r = 0.38 and a p = 0.04. Correlation of the percent change in the ΣQRSs between admission and points of the peak body weights, with the percent change in the corresponding albumin values, revealed an r = 0.37 and a p = 0.053.
The difference in our results from the admission time point and the time of peak weight is hard to explain. If there was a realrelationship between QRS amplitude and albumin, such a discrepancy should not have been encountered. Moreover, I find it odd to correlate absolute values of albumin with absolute values of amplitude of QRS complexes, taking into consideration that the latter can vary widely among both healthy subjects and patients, while the normal range of the former is 3.3 to 5.3 g/dl in our laboratory.
I find the correlation of change in each of the two correlates more meaningful scientifically, and in our data we found an almost significant association between the drop of albumin and loss of QRS amplitudes. The correlation coefficient was much smaller than the one we reported (r = 0.61) (2)for the association of weight gain and ΣQRS drop. However, some association may exist between albumin and QRS amplitudes, and may be confirmed in patients in whom both daily electrocardiograms and albumin measurements are implemented.
Does this mean that the hypoalbuminemia was causally related to the drop of ΣQRS in our patients, or was this biochemical abnormality “an innocent bystander”? Innocent it is not, for certainly it contributes to the extent of water retention; however, the causative mechanism for the drop in ΣQRSs must have been the weight gain (or fluid retention). Correlation of % weight gain and % albumin drop in our 28 patients was poor (r = 0.02, p = 0.9); also, correlation of the % weight loss and % albumin rise in 9 patients was poor (r = 0.5, p = 0.1). It would have been very interesting to have data on weight change in the patients reported by Heaf (1).
There is evidence that edema fluid has low resistivity (3); thus, its effect on the transfer of cardiac potentials from the heart to the body surface can be explained by Ohm’s law. I am not aware of any work ascribing any mechanistic role for albumin per se in modulating the ECG voltage. Finally, the fact that patient 26 (2)gained 122.4 lb and lost 68.3% of his ΣQRS, while his albumin increasedfrom 1.8 g/dl to 2.7 g/dl, and subsequently lost 77.4 lb and gained 185.9% of his ΣQRS (Fig. 6 of Madias et al. ) while his albumin increased merelyfrom 2.9 to 3.0 g/dl, casts serious doubt on the contention that a causative relation between albumin and QRS amplitudes really exists.
- American College of Cardiology