In this study, we showed that BD is used commonly by anesthesiologists as an endpoint of resuscitation ( > 70% of respondents use BD at least "sometimes"). The absolute value of BD used by different clinicians for assessing was variable, and many clinicians indicated that the absolute BD value was not as important as the trend. In comparison, fewer of these same respondents used serum lactate as an endpoint of resuscitation. In this study, we have demonstrated that BD is not a suitable surrogate for hyperlactatemia. These data are consistent with our previous findings in critically ill patients[11
]. Moreover, our data indicate that the base deficit would often mislead the clinician as to the true serum lactate. We showed that the ROC area under the curve for BD to predict hyperlactatemia was 0.58 (0.50 is a coin toss); thus demonstrating the poor performance of BD as a surrogate for serum lactate (Figure ).
These results are not surprising given previous literature assessing the capacity of base deficit as a surrogate for resuscitation and serum lactate in the critical care and trauma setting. We have previously shown that BD is not reliable as a surrogate for serum lactate in critically ill patients[11
]. In addition BD has also been shown to be insensitive for detecting hyperlactatemia in surgical patients[18
]. In a study by Mikulaschek et al of trauma patients, resuscitation decisions would have been wrong 33-58% of the time if BD or anion gap had been used as the sole criterion rather than serum lactate concentration[16
]. Waters et al showed that in a study of surgical patients, elevations in BD without concomitant rise in lactate were attributed to hyperchloremia and were a manifestation of successful resuscitation rather than fluid deficit[19
]. Base deficit is useful as a surrogate for resuscitation so long as it is a result of hyperlactatemia, but if it is caused by a hyperchloremic acidosis due to saline loading this relationship becomes discongrous[10
]. In addition, base deficit as a surrogate for serum lactate can be confounded by renal dysfunction and ketoacidosis.
This study is the first to test this concept on patients undergoing general anesthesia in the operating room. These data are noteworthy because based on our survey results, clinicians routinely use BD as an endpoint of resuscitation. The assumptions underlying this physiology may be faulty in many patients because BD does not discriminate the cause of the acidosis. Since previous studies have shown that the accumulation of chloride from normal saline tends to increase the base deficit[10
], many patients in the peri-operative setting may receive more volume and blood product transfusions than they should because the BD is increasing due to the saline loading, and the clinician may incorrectly interpret this rising BD as an indication that the patient in under resuscitated. In this study, we were unable to show that the discrepancy between BD and serum lactate worsened with the amount of volume resuscitation or the duration of the case (Figure ). This is not surprising given the multitude of intra-operative activities that can affect acid base status ( e.g. resuscitation, saline from iv fluids, citrate from blood products, etc) However, since the level of agreement between BD and serum lactate was so poor at baseline, the effect would have had to be dramatic in order to detect it.