The Reply

Abstract

We thank Xue et al for their interest in our article in which we described the association between preoperative hypernatremia and 30-day perioperative morbidity and mortality. In their letter, they raise several concerns relating to some aspects of our study. First, Xue et al point out that sodium levels are subject to change over time. Indeed, blood work drawn far in advance may not necessarily be reflective of sodium levels on the day of surgery. Although in our primary analysis, we included sodium levels collected as far as 1 month before surgery, we also performed several sensitivity analyses to account for changes in sodium over time.1Leung A.A. McAlister F.A. Finlayson S.R. Bates D.W. Preoperative hypernatremia predicts increased perioperative morbidity and mortality.Am J Med. 2013; 126: 877-886Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar Our findings were robust, and the association between preoperative hypernatremia and 30-day perioperative mortality remained elevated irrespective of when sodium was collected (adjusted odds ratio [aOR], 1.48; 95% confidence interval [CI], 1.36-1.61 for within 7 days; aOR, 1.55; 95% CI, 1.42-1.69 for within 24 hours). In fact, the magnitude of association was greatest for blood work collected within the same day of surgery, suggesting that our overall results were actually conservative estimates of true risk. Second, we agree that an inherent limitation of all observational studies is that unmeasured confounding can never be excluded confidently despite careful and appropriate adjustment for differences between groups. However, we believe that our study provides the best possible evidence given the nature of the subject. We disagree with Xue et al that randomized controlled trials are needed to confirm the association. Although it is true that a well-conducted randomized controlled trial is considered the gold standard for questions of therapeutic efficacy, it may be altogether inappropriate for questions of prognosis or natural history where exposures cannot be practically controlled or ethically conducted (ie, it is infeasible to induce hypernatremia in otherwise healthy subjects before surgery). Third, although we had access to rich, validated, clinical information on many important perioperative risk factors, we acknowledge that we did not have detailed data on intraoperative anesthesia. Nonetheless, we demonstrated that the general association between hypernatremia and perioperative mortality was consistent for inpatients and outpatients, emergency and nonemergency cases, as well as across a broad range of surgical conditions, thus suggesting that the observed association is less likely a sole product of differences in anesthetic management. Finally, we recognize that preoperative anemia also is associated with 30-day perioperative morbidity and mortality.2Musallam K.M. Tamim H.M. Richards T. et al.Preoperative anaemia and postoperative outcomes in non-cardiac surgery: a retrospective cohort study.Lancet. 2011; 378: 1396-1407Abstract Full Text Full Text PDF PubMed Scopus (722) Google Scholar Accordingly, we have now re-performed our analysis using preoperative hematocrit levels, defining anemia as a hematocrit level less than 36.0% in women and less than 39.0% in men, as in previous studies.2Musallam K.M. Tamim H.M. Richards T. et al.Preoperative anaemia and postoperative outcomes in non-cardiac surgery: a retrospective cohort study.Lancet. 2011; 378: 1396-1407Abstract Full Text Full Text PDF PubMed Scopus (722) Google Scholar Of note, we found that the inclusion of this variable did not meaningfully change the association between preoperative hypernatremia and 30-day perioperative mortality (aOR, 1.44; 95% CI, 1.33-1.56 when hematocrit levels were excluded; aOR, 1.43; 95% CI, 1.32-1.55 when hematocrit levels were included). Is Preoperative Hypernatremia an Independent Predictor of Perioperative Morbidity and Mortality?The American Journal of MedicineVol. 127Issue 1PreviewIn a recent study, Leung et al1 concluded that preoperative hypernatremia was associated with increased perioperative 30-day morbidity and mortality. The power of this study is its use of a large dataset from the American College of Surgeons National Surgical Quality Improvement Program, which includes and adjusts for most of the known risk factors affecting perioperative morbidity and mortality. Furthermore, they used appropriate methods to assess associations between hypernatremia and observed end points. Full-Text PDF