Knowledge of factors that affect or prevent microvascular thrombosis is crucial to surgeons performing such operations. Although some studies have evaluated the potential protective role of anticoagulation medication in microvascular thrombosis, there is yet no consensus about their effectiveness in preventing such complication.
We read with interest the article by Lee et al.1 and would like to comment on their retrospective study investigating the effects of ketorolac on microvascular thrombosis in lower extremity reconstruction. The authors criticize previous studies suffering from methodologic flaws and material heterogeneity. Although they praise the relative homogeneity of their material (four types of flaps) that would make their study more valuable than others, we struggle to understand the meaning of “relatively homogeneous,” as statistical material can be considered either homogeneous or heterogeneous.
Even though they found no significant differences between procedures, we believe that each type of flap is related to a particular technical complexity with specific operative time, and even the same surgeon may have different skills in harvesting various flaps. As a result, each procedure has an inherent risk of vascular-related complications, and grouping heterogeneous flaps could increase the probability of misinterpreted relationships between ketorolac and the outcome categories. Moreover, the authors only report data on a number of eventful outcomes, but not the frequencies in each flap type; this information is fundamental for better clarification of outcomes variability and sampling bias.
From their analysis, the authors proved a significant relationship between ketorolac and two outcome categories: minor complications (17 events) and total complications (19 events). Nevertheless, we observed that the latter outcome results from the sum of minor and major complications (two flap losses). As a consequence, the outcome total complications emerges strongly dependent on minor complications, and its significant relationship to ketorolac results is biased by such outcome selection. This basic methodologic flaw could persuade the reader about the effectiveness of postoperative ketorolac medication in microvascular thrombosis prevention, as the authors suggested in their conclusions, potentially resulting in a stronger inclination to treat patients addressed to microsurgery with ketorolac.
Many studies2 have widely shown that the method with which quantitative data are selected and summarized can have great effects on the likely interpretation of those data, causing subsequent differences in care. Given our considerations of the authors’ study design, it is important that multivariate analyses explore both large enough overall sample size and equal groups of subjects, regarding adequate subject-to-variable ratios, while performing various checks of internal and external validity, or cross-validation.3
However, regression analyses from retrospective studies, such as that reported by Lee et al., can allow the reader to judge solely the magnitude of an association, but do not express the clinical implications as clearly as controlled clinical trials. This distinction can create deep divergences in clinical decision-making. We believe that ketorolac is a very effective drug for the treatment of moderate to severe pain, but its hypothesized efficacy in microvascular thrombosis prevention still has to be carefully evaluated, as there are no definitive data to resolve such controversy.
The authors have no financial interest to declare in relation to the content of this communication.
Benedetto Longo, M.D.
Fabio Santanelli, M.D., Ph.D.
“Sapienza” University of Rome
School of Medicine and Psychology
Plastic Surgery Unit
1. Lee KT, Jeon BJ, Lim SY. The effects of ketorolac on microvascular thrombosis in lower extremity reconstruction. Plast Reconstr Surg. 2012;129:1322–1327
2. Forrow L, Taylor WC, Arnold RM. Absolutely relative: How research results are summarized can affect treatment decisions. Am J Med. 1992;92:121–124
3. Albert A, Harris EK Multivariate Interpretation of Clinical Laboratory Data.. 1987 New York Dekker