Now, the question arises: Would this higher failure rate (still) be acceptable in the context of fewer CTPA tests in these elderly patients? Basically, this comes back to answering the following two questions. First, how many additional PE cases are missed, and how many of them are fatal? Second, how many CTPA tests are avoided? This is important as at least in some patients CTPA will be performed without PE being present, and these patients are at risk for unnecessary iatrogenic damage from CTPA (including radiation and contrast nephropathy). Ideally, these questions should be addressed in a formal randomized controlled trial (RCT) comparing both diagnostic strategies. However, RCTs comparing diagnostic strategies are in general difficult to perform (as well as to finance) and for now are lacking. Indirect comparisons may, however, be more feasible. For instance, if we compare aggregated data from an also recently published prospective evaluation of the age-adjusted cutoff (the ADJUST-PE study7), plus the study by Woller et al,4 with a landmark and large-scale diagnostic management study in suspected PE using a conventional cutoff (the Christopher Study8), the following conclusions can be drawn (Table 1). First, using either a conventional or an age-adjusted cutoff, the proportion of missed PE cases is comparable and small (well below 2%), and no fatal cases occur, which is a reassuring finding. Second, more strikingly, the proportion of patients for whom CTPA is still required is also rather comparable: Still around six to seven in every 10 suspected patients require CTPA to confirm or refute a diagnosis of PE, a finding in concurrence with a systematic review on decision rules in suspected PE.9 This leaves us startled. Comparisons within cohorts demonstrate a sharp reduction of required CTPA tests, notably in the eldest elderly. Yet, an indirect comparison over different cohorts does not demonstrate this reduction of CTPA tests. Obviously, inherent problems related to this indirect (nonrandomized) comparison play a role here. Combing individual patient data (IPD) of these studies and using state-of-the-art IPD meta-analysis10 could allow us to explain these different findings between within and over cohort comparisons. For instance, the number of comorbidities, the applied clinical decision rule, the prevalence of PE in different studies, increasing age, D-dimer assay used, as well as the applied reference standard (as over time CTPA has become increasingly more sensitive in finding smaller emboli) could all be evaluated as potential modifiers of the safety and efficiency of this age-adjusted D-dimer. This would provide us valuable information on the subgroup of patients that benefit most from using this age-adjusted cutoff.