Changes in Intrapulmonary Shunting

We would like to offer the following comments concerning the interesting paper by Shapiro et al (Chest 1980; 77:138). Even though we agree, in general terms, with their statement that venous admixture ( va/Qt) should increase when a patient breathes 100 percent 02, measured either by Berggien’s classic methodl or with the inert gases technique,2 we have observed that it does not always do so because the increase depends upon multiple factors. When Qva/Qt is measured breathing room air, the final value will reflect gas exchange abnormalities caused either by VA/Q inequalities, diffusion impairment or shunt (Qs/Qt); and if F1o2 is gradually increased reaching 100 percent, final findings will depond upon the interrelationship between the correction of the first two factors and the eventual occurrence of true shunt that ventilation with 100 percent 02 usually provokes. That explains the significant correction of QVa/Qt in conditions with only moderate inequalities of A/Q when F102 is increased slightly. Besides, alterations of gas exchange produced by VA/Q inequalities at F1o2 greater than 50 percent do not play an important role.2 At this F1o2 level, Qva/Qt modifications are possibly due to very important A/Q mismatching (log normal distribution of VA/C, log SD(u)> 1.5)2 and/or eventual diffusion impairment with concomitantly increasing shunt. It seems clear to us that in patients with slight A/Q inequalities and no diffusion impairment F1o2 increments will enhance true shunt caused by the absorption atelectasis that high 02 administration would produce. the other hand, in cases with severe A/Q inequalities and eventual diffusion impairment. the correction of these defects can counterbalance the shunt increment caused by a high F1o2 and Qva/Qt decrease accordingly. In this context it is interesting to report our results in 16 patients with bacterial bronchopneumonia who were mechanically ventilated with different F1o2 ( Fig 1 ). When Qva/Qt with maintenance F1o2 was compared with the Qva/Qt on pure oxygen, no significant differences were found. If the patients were divided in two groups (those with Qva/Qt less than 0.20, and those with Qva/Qt >0.20), we observed that in the former group Qva/Qt did not change significantly when on 100 percent oxygen (0.16 vs 0.17, n =7), while it decreased significantly in the other group (0.57 vs 0.41, P< 0.05, n = 9). These findings demonstrate the existence of two different populationswith dissimilar Qva/Qt response to F1o2 increments. These Qva/Qt reductions on F1o2 increments have been reported earlier by other authors. The above mentioned data do not agree with the conclusion of Shapiro et oil that “the original intent of administration of 100 percent oxygen . . . is not valid since . . . results in an increase in the calculated shunt.” This assertion is possibly true in their particular group of patients with moderate alterations in gas exchange ( Qva/Qt =0.17), but it is not applicable at all to more severely affected patients, eg, Qva/ Qt> 0.40 in which the relationship of altered mechanisms of gas exchange are probably different. We also disagree with thefr statement, “We feel confident that there is no clinical advantage to making this measurement at an F1o2 of 1.0.” In this context we find it very interesting that King’ when referring to patients with catastrophic pulmonary failure, points to the usefulness of measuring shunt modifications with different FIo2. In the case of no change of shunt when F1o2 IS modified, this could be reduced withouta significant decrease of Pa02. On the contrary, if Qs/Qt decreases 0.70 1-