SEQUENTIAL-CHANGES IN VITAL SIGNS AND ACID-BASE AND BLOOD-GAS PROFILES IN PNEUMOCYSTIS-CARINII PNEUMONITIS IN CHILDREN WITH CANCER - BASIS FOR A SCORING SYSTEM TO IDENTIFY PATIENTS WHO WILL REQUIRE VENTILATORYSUPPORT

Early reliable identification of patients with Pneumocystis carinii pn eumonia (PCP) who will require ventilatory support would be desirable. To develop a predictive system to meet this need, we studied, prospec tively, the sequential alterations in vital signs and acid-base and bl ood-gas profiles associated with this disease in 55 children with canc er, 29 of whom did not require ventilatory support (Group I) and 26 wh o did (Group II). None of the patients had acquired immunodeficiency s yndrome (AIDS). On admission to the hospital the only feature that dis tinguished patients in Group I from those in Group II was the mean (+/ - SD) respiratory rate (38.7 +/- 2.1 versus 49.1 +/- 3.5 breaths/min, p < 0.02). By 12 h after admission there was a significant difference in the partial pressure of oxygen (Pa-O2) between Groups I and II (75. 1 +/- 3.2 mg Hg versus 65.4 +/- 3.1 mm Hg, p < 0.05), and also in the two groups' inspired fraction of oxygen (Fl(O2); 24.9 +/- 0.54% versus 29.6 +/- 1.6%, p < 0.01). Both alterations, as well as tachypnea, per sisted for the remainder of the study period. The maximum Fl(O2) did n ot exceed 45% in Group I, and by 60 h after admission to the hospital, all patients in this group had persistent increases in Pa-O2 that exc eeded 80 mm Hg, permitting decreases in Fl(O2) to that of room air. In Group II, hypoxemia was refractory despite an increase in Fl(O2) to 5 0%, at which point venti[atory support was begun (at a mean of 81.1 +/ - 32.3 h after admission). Discriminant-function analysis of the vital signs and blood-gas and acid-base measurements in these patients allo wed us to devise a scoring system to predict the need for ventilatory support during the course of PCP, Scores based on information collecte d at 36 h after admission would have identified with 91% success those patients who required ventilatory support. These findings demonstrate distinctive alterations in respiratory rate and arterial oxygenation associated with PCP in children with cancer without AIDS who subsequen tly require ventilatory support. The proposed scoring system, after va lidation in other patient cohorts with PCP associated with either AIDS or other immunosuppressive diseases, may permit more timely decisions regarding early hospitalization and supportive therapy, which may be lifesaving.