Letter to the editor

From a methodological point of view, there is little description of controls, including no description of whether they were identified concurrently at the time of case enrollment or retrospectively, whether they were selected from all hospitalized patients or a subset, or how the authors determined that controls did not have evidence of pneumococcal disease (presumably not by prospectively testing each patient). Since the key source of bias in a case control study often is control selection, absence of these data makes assessment of potential bias impossible. [...]unlike PPSV23, PCV13 consistently has demonstrated VE against NBPP in a randomized controlled trial [3] and real-world study [4], has relatively high direct VE against serotype 3 [5], provides at least four years of undiminished protection to adults [6], and is effective among adults > age 74 years [7]. [...]given all the above, we think that the authors conclusions in the current manuscript might be qualified further.