CORRESPONDENCE in homozygous sickle cell disease (SS) and 5.5 percent in SC/S-/3-thalassemia disease. The numbers of patients were 26 and 19, respectively. The difference in frequency between the two groups is not statistically significant (P>0.05). Yet the authors state that patients with Hb SC and Hb S-fl-thalassemia only should arrange in advance for inflight oxygen. They add: "We believe that adult patients with Hb SS do not need supplemental oxygen unless they are among the 6 percent of adult patients with Hb SS and intact spleens." Admittedly, there are fewer instances of splenic crises in SS disease (4.3 percent as compared with 8 percent in other forms), but it is not splenic crises only that we need to or can prevent by supplemental oxygen. The 6.5 percent of patients with Hb SS who do suffer vasoocclusive crises could benefit from inflight oxygen. The authors also found that whereas 42.9 percent of SS patients in Reno (elevation 4,400 ft) suffered vasoocclusive crises, only 31 percent of SS patients in Lake Tahoe (elevation 6,320 ft) had similar complications. This difference is significant (P<0.05). Watson-Williams also noted this in his critique of the paper,2 but neither he nor the original authors could provide an explanation for this seeming contradiction. In fact, there is no contradiction. Among other things, the sickling process of deoxylfib S is allosterically modified by pH and 2,3-diphosphoglycerate (2,3-DPG). Lowering the pH hastens sickling by decreasing the oxygen affinity of Hb S via the Bohr effect. In contrast, 2,3-DPG has no independent effect on the polymerization of Hb S. Instead it alters pH, which in turn enhances polymerization.3 It follows, therefore, that if the pH effect of 2,3-DPG is annulled by respiratory alkalosis, which occurs to a greater degree at 6,300 ft than 4,400 ft, it cannot influence polymerization of Hb S. While this phenomenon cannot adequately explain the reduced frequency of crises in patients from Denver when they are exposed to high altitudes,4 a comparison of the two studies done so far on the subject" 4 cannot be made since one recorded the frequency of vasoocclusive events' and the other recorded the number of patients who developed similar symptoms.4 In any event, this is one of the more intriguing findings in Claster and associates' study and deserves future investigation.