Our study has several limitations. The number of patients who had severe anemia (Hgb level, <7 g/dL) and were not transfused is relatively small compared with the numbers in the other groups used for comparison. Because our study was retrospective, we could only find associations and not prove causation. This is significant, as the higher complication rate seen with transfusions may only be caused by the transfusion as a predictor of a patient requiring more complex surgery with higher blood loss (and higher risk of complication) or other such risk factors that led to transfusion, but not the transfusion itself causing the complication. An attempt was made to remove this potential bias by controlling for age, sex, ISS, and whether the patient had multiple surgeries. However, there may have been other significant confounding variables not excluded. As complications were assessed by chart review, they may not include those that occurred at other institutions and that were never reported to the practitioners at our facility (though we did have the ability to search records of neighboring institutions electronically when electronic medical records were available). That no functional outcomes were included in this retrospective review might make the complication rate appear more or less sensitive than the patients’ own opinions regarding their outcomes. All these weaknesses could call into question whether the statistically significant higher risk associated with allogeneic transfusion found in this study is real, but the focus and reason for pursuing this study were to determine if permissive anemia was dangerous or would be associated with a higher risk of complications than routine allogeneic transfusion of asymptomatic patients to treat a laboratory value.
Strengths of the study include the review of a single surgeon’s practice with a written protocol in place for anemic orthopedic trauma patients. The 95% follow-up (104/109 patients) is good for this type of trauma population. Although this series is retrospective, it is reasonably large for a subgroup of young, healthy orthopedic trauma patients to study the effects of anemia or transfusion. Whether transfused or not, many of these patients tolerated Hgb levels under 7 g/dL, which gave a large enough comparison group to evaluate the independent effects of transfusion (or of using transfusion as a marker for complication risk) or anemia as a risk factor. As a result, it appears that a more conservative transfusion strategy may be as safe as a more liberal transfusion strategy. The results of this retrospective study were used to design a prospective multidisciplinary pilot study randomizing patients to either a liberal or a conservative transfusion strategy to determine which approach might carry higher risks of complications.
Conclusion
The results of this retrospective study suggest that a conservative transfusion strategy in a young, healthy, euvolemic asymptomatic patient who is not actively bleeding may be as safe as a liberal transfusion strategy and potentially may have fewer complications than does transfusion for a conventional laboratory value. Our study results do not suggest that transfusions should be held in patients who are symptomatic at rest or in patients who are being actively resuscitated, as transfusion can be lifesaving under these circumstances. A prospective randomized study has begun at our institution with enrollment expected to take 2 years with another year needed to complete 1-year follow-up of all patients.