Sometimes, it can be hard to get a handle on what really big numbers mean. If we say that only about one in a million patients who gets a vaccine injection of any sort will have a problem severe enough that the US National Vaccine Injury Compensation Program will remunerate that patient, it seems like a small problem. But if we learn that the program has paid out over USD 4.7 billion, most of that in the last 10 years [16], we start seeking a sense of scale. UNICEF says it costs between USD 37 and USD 101 to fully vaccinate a child against 11 diseases through the age of 24 months [37]. Given that, USD 4.7 billion would cover the vaccine costs of somewhere between 46 million and 127 million children. Of course, if the vaccines genuinely caused injuries, then fairness suggests that the money was well spent right where it was. It’s worth noting that the compensation program protects healthcare practitioners who give the vaccines as well as the pharmaceutical industry that develops the vaccines, and this may help encourage development and delivery of vaccines, which is—as we’ve seen lately—a pretty important goal. This matters to orthopaedic surgeons for another reason, though: Although some compensation for complications following vaccination pertains to severe allergic reactions, much has gone to patients who reported persistent shoulder symptoms after receiving a shot in the arm. This should cause us to ask: Do vaccines cause shoulder injury? The National Academy of Medicine says yes [22], and a position statement from the American Academy of Orthopaedic Surgeons says no [1]. Sounds like a horse race. In this month’s Clinical Orthopaedics and Related Research®, two systematic reviews likewise have two different takes on the matter. One, from Dell Medical School at the University of Texas at Austin, under the guidance of David Ring MD, PhD [33] makes a compelling case that the types of shoulder symptoms that follow vaccine administration are so common even in the absence of any shots that any such symptoms in this setting are almost certainly coincidental. This is a point that his team has explored before using other methods, with similar results [15]. They emphasize that promoting the idea that vaccines can cause this kind of harm is apt to contribute to vaccine hesitancy, which is something we can ill afford as we try to emerge from the pall of COVID-19. The other systematic review, from a team at Johns Hopkins University lead by Uma Srikumaran MD, MBA, MPH [25], suggests that discounting the possibility that vaccines can in rare circumstances cause harm, when we don’t know for sure, may result in the very same problem. Either way, it's all about trust, and when we give it away, it’s hard to get back.Fig. 1: Uma Srikumaran MD, MBA, MPHFig. 2: David Ring MD, PhDIt’s so rare that I see such starkly different takes on the same topic that I felt the need to run these manuscripts side by side and bring the two senior authors together for a conversation. Read their papers in this issue of CORR® [25, 33], and then join me as I try to sort out the question of “do they or don’t they,” with Drs. Ring and Srikumaran in the Take 5 interview that follows. We will try to convince one another whether shoulder pain—or is it “shoulder injury”?—after vaccine administration is a diagnosis or not. Take 5 Interview with David Ring MD, PhD and Uma Srikumaran MD, MBA, MPH, senior authors of two systematic reviews with opposing conclusions on shoulder pain after vaccine administration. Seth S. Leopold MD:Persistent symptoms in the shoulder after a vaccination are extremely rare, and the associated diagnoses include a host of things that either occur incidentally with great frequency (such as the spectrum of rotator cuff tendinopathy, which may or may not include a component of “bursitis”) or seem unlikely to be causally linked to the jab itself (like osteonecrosis of the humeral head). It’s easy to describe the overlap as coincidental, and I see why the post hoc ergo propter hoc fallacy is tempting. But isn’t it possible that there’s not just smoke, but also fire? And if so, shouldn’t we remain open to the possibility? David Ring MD, PhD: Remaining open to the possibility and promoting the possibility—intentionally or inadvertently—are two different things. I’m open to the possibility that playing my guitar too much or typing out these articles so often can damage my arms. I’m also aware that the belief that activity causes harm is problematic, as witnessed in the human-made epidemic of so-called “repetitive strain injury,” the remnants of which continue to harm people to this day [3]. I take great care to interpret discomfort with cherished activities in the most healthful manner. One of my mentors, Arthur Barsky—an expert in the psychosomatic—wrote an editorial entitled, “The Iatrogenic Potential of the Physician’s Words” [4], in which he details how what we say and do can have important influences on the health of individuals and society. Before we theorize that “vaccination can harm the shoulder,” we should first consider the potential for harm inherent in the concept. Theories, injury, or harm can reinforce our natural human tendencies to interpret a new pain as signaling a new pathology [23, 38] or perhaps an injury [12, 24], and a painful activity that will make the problem worse [8, 34, 39]. These natural—albeit, largely unhelpful—thoughts are measured using instruments such as the Pain Catastrophizing Scale [34], the Tampa Scale of Kinesiophobia [39], and the Negative Pain Thoughts Questionnaire [8]. Unhelpful thinking accounts for a substantial portion of the variation in comfort and capability among people with a variety of musculoskeletal conditions, with the severity of pathology having much less influence [7, 12, 28]. It’s going to be extremely difficult to identify a vaccine-specific pathology. If we can achieve that, then we’d need to prove that this vaccine-specific pathology causes permanent harm. Shoulder pain after vaccination is likely common, not rare [5, 15]. Given that most of us develop rotator cuff tendinopathy as we age [35], and many of us get an annual influenza vaccination (among others), the probability that some people will first notice symptoms from their established rotator cuff tendinopathy after an injection in the shoulder is high. What’s thankfully rare—for the time being—is a person with new persistent symptoms misperceiving those symptoms as an injury [12, 23, 38], and also having that misperception reinforced and codified by healthcare professionals, something that is our responsibility to prevent. All signs point to erring on the side of concluding there is no harm to the shoulder from vaccination in the absence of compelling experimental evidence to the contrary. Until there is reproducible, experimental evidence of a vaccine-specific pathology, we can be curious about whether one exists, but we should be mindful of the potential for substantial psychological, iatrogenic, and financial harm to individuals and society if we promote the concept in the absence of such evidence. Dr. Leopold:Dr. Srikumaran, what’s wrong with that answer? Uma Srikumaran MD, MBA, MPH: We must remember that absence of evidence is not the same thing as evidence of absence. I think we disagree on how to interpret the best-available experimental evidence to date. I believe two articles [18, 20], representing Level III data, provide some of the experimental data that Dr. Ring seeks. Interestingly, this evidence is congruent with the Level IV and V evidence over the past decade as well as my own experience. The research indicates that the additional risk of bursitis specifically related to vaccination over the baseline risk is common, as Dr. Ring accurately notes [18]. But Hesse et al. [18] looked at that baseline level of bursitis in a control group and showed that the additional risk of bursitis—time-linked to vaccination—is 1:150,000 over the quite common baseline bursitis that exists in the general population not related to vaccination. And so, according to the best-available evidence, Shoulder Injury Related to Vaccine Administration (SIRVA) is, in fact, rare [18]. Dr. William Osler noted: “Medicine is a science of uncertainty and an art of probability.” He also said: “Listen to your patient—he is telling you the diagnosis.” I am inclined to believe that our greatest responsibility is listening to our patients and guiding them toward treatment (and educating on prevention for the future). I think we agree we can acknowledge the patient experience, reorient the focus of patient discussion to the readily treatable nature of all types of bursitis, and focus on future preventive strategies, thereby mitigating the “harmful” thoughts patients may have, which is exactly what Dr. Barsky [4] advocates for in his article. I outline the harms of not acknowledging the patient experience and the current evidence in my response to the question about harms below. Dr. Leopold:Dr. Srikumaran, the fact that you call it SIRVA means that on some level you believe there is not just a plausible but also a likely causal link between the shots and the symptoms that follow, or at least you’re pretty receptive to the possibility. Given how soft the supporting evidence for causality is, how do you get there? Can you use something like the Bradford Hill criteria[19]to make the case, or do you do it some other way? Dr. Srikumaran: I can appreciate that the term SIRVA itself is problematic. What does “injury” mean exactly? I believe that the “I” in “SIRVA” would be more accurate if it meant “inflammation” rather than “injury.” If we focus on “inflammation” and not chronic disorders of the shoulder (labral or rotator cuff tears, arthritis) I think we can appreciate how a vaccine inadvertently injected in or near the bursa, a structure known to become inflamed, might lead to a painful shoulder and indeed be plausible. As is noted in both reviews, the concept is relatively new, so of course the science has yet to catch up to provide definitive evidence. Like most practitioners who may have seen a patient present with persistent shoulder pain consistent with bursitis after receiving a vaccine, I initially was skeptical and thought it might be coincidental. However, as a shoulder specialist, I began to see a pattern among my own patients—they tended to be quite thin, often reported a high injection site, and were quite convincing in their reports of the pain and associated symptoms. Most of these patients were quite familiar with the typical muscular pain and soreness after a vaccination into the deltoid, as it is a common yearly event. Those who presented with a picture that struck me as plausible for SIRVA could clearly characterize the difference in the pattern, timing, and progression of pain, and the majority had no clinical history of chronic shoulder conditions (of course any type of imaging at that time will reveal all kinds of chronic conditions, it is important to remember these were completely quiescent in most of these patients). Further, these patients were not pursing litigation nor were they aware of the federal compensation program. I also believe the theory of SIRVA holds up well in terms of Bradford Hill’s criteria for causation. As SIRVA is a rare condition, and the evidence largely limited to case series, these criteria can help suggest if an association might have a causal relationship. Atanasoff et al. [2] considered these in their discussion of the topic, and I think the evidence has only grown since then. The first criterion is plausibility, which we covered earlier. Other criteria include specificity, temporal association, consistency, strength (dose-response relationship or reversibility/risk factor relationship), and experimental evidence. Most patients with SIRVA received a vaccine with a rapid onset of symptoms (temporality), no history of shoulder pain or dysfunction prior, with symptoms isolated to the area of injection (specificity), with physical findings consistent with a local inflammatory injury (plausibility), and many recalling a mechanism—the high-deltoid injection—that could result in injection of antigenic material into the bursa (which is a risk factor relationship—the higher incidence with a reported higher injection site). It is important to note that simple needle injection without antigenic material with other agents, such as lidocaine or platelet-rich plasma (PRP), has not been associated with shoulder injuries. This lends further support and reliability to the theory that antigenic material (not simply any type of material or simply needle placement alone) is the likely cause of possible immune-mediated inflammation (specificity criterion of causation). The criterion of reversibility is also established in that if the proposed cause is removed (antigenic material), the effect (SIRVA) is not seen (again, there are no reports of similar issues with lidocaine and other injected materials). Further adding to the criterion of reversibility are the recent reports of successful treatment with steroid injections [26] with resolution of symptoms within 1 month after corticosteroid injection into the subacromial bursa within 5 days of receiving the vaccination. Considering that SIRVA is thought to be an inflammatory and immune response within this bursal structure, the fact that a strong anti-inflammatory delivered to the same structure can resolve it is important evidence that contributes to the prevailing theory via the criteria of reversibility. The criterion of consistency is also supported as independent researchers from separate institutions, different continents, different sample sizes, and different time periods (2007-2021) have reported similar experiences. Finally, experimental evidence in a Level III comparative study now exists [20], adding to the Level III epidemiologic comparative study by Hesse [18] that suggested an incidence of 1 in approximately 150,000. The Hirsiger study [20] investigated vaccine-related shoulder pain in a group of students vaccinated with the quadrivalent influenza vaccine. All of the students with pain (except for one) were vaccinated in the same medical student vaccination campaign in 2018. The control group was recruited from a separate, prospective vaccination cohort and was then matched for the interval between vaccination and blood sampling. When compared to the control group, those with imaging findings were found to have higher levels of immune markers associated with influenza vaccination (Tfh-like T cell subsets, activated B cells) that persisted in the peripheral blood for several months. This not only provides an explanation for immune-mediated tissue response, but also why patients often have ongoing symptoms for several months after vaccination despite treatment. Dr. Leopold:Dr. Ring, what’s wrong with his reasoning here? Dr. Ring: We probably wouldn’t be talking about shoulder pain after vaccine administration if it was conceptualized as harmless inflammation comparable to what one routinely experiences in the deltoid muscle after vaccination. The concept of “SIRVA” implies permanent, compensable harm (“injury”). Dr. Srikumaran and I agree that there is no evidence of permanent, unique, vaccine-related harm even in the most closely studied patients. The word “injury” is likely being misapplied [5, 38] as words often are in discourses about musculoskeletal health [32]. Humans invented the scientific method because human intelligence is prone to false associations [31]. A key aspect of the scientific method is the assumption of no relationship: the null hypothesis [29]. Null hypotheses are particularly important with speculative causal associations. There is a notable history of harm caused by speculative associations. In the vaccine realm, what comes to mind is the fraudulent association of vaccination and autism. Such false associations capture our imagination and are difficult to undo [21]. Another is the famed false association between proximity to power lines and childhood leukemia [10]. It was determined that the observed association resulted from a form of confirmation bias known as the Texas Sharpshooter Fallacy: when similarities in data are overemphasized and differences in data ignored [13]. This form of cognitive bias is named after a joke about a Texan who shot a bunch of rounds at the side of a barn then drew a bullseye around the largest cluster of bullet holes and claimed to a be a sharpshooter. The association between cancer and power lines was ultimately disproved with better population-based data [6]. Given the potential harms associated with claims of causation, humans should assume no association and no harm until a high burden of experimental proof is met. Dr. Srikumaran’s description of his journey to believing that vaccines can injure the shoulder can serve as an instruction manual for how cognitive bias leads to false association. First, anchor on the idea, then confirm it through selective attention. Next, experience reinforcement of the theory from sympathy for the patient and desire to do the right thing and advocate on their behalf. Add in a little stress contagion (adopting the patient’s explanatory model [“this one is different”] and concerns [“something’s wrong, the shot was too high”]). Ideas can seem compelling while also being incorrect or unhelpful [14]. It’s possible to validate a person’s experience while simultaneously considering his or her interpretations with a critical, scientific eye [9]. Indeed, this is what the caring, compassionate physician must do. Misinterpretation of symptoms is extremely common [8, 23, 38]. In one recent study, about 1 in 5 people misperceived new symptoms from an established, gradual onset pathology as representing new pathology, and it often seemed like an injury [23]. In another study, 37 of 119 (31%) patients with hip and knee osteoarthritis perceived their problem as an injury [12]. It is the physician’s duty to anticipate and identify common misconceptions before they cause harm. The Bradford Hill criteria [19] cannot be satisfied by the rationale as Dr. Srikumaran offers, and there is inadequate experimental evidence for any of the criteria to date. New data from Hirsiger and colleagues’ study of 16 medical students (much younger than the typical patient with SIRVA) with persistent shoulder pain after exposure to relatively inexperienced injectors (not clear who—perhaps the students themselves) identified four with bone erosions that they thought were distinct from the cystic changes that are common near the enthesis posteriorly, unrelated to age or variations in rotator cuff tendon pathology [11]. The four students with erosions had elevation of specific immune responses similar to those seen with intentional intraarticular injection [36]. While the authors note several important limitations to their work that prevent them from concluding definitively that these bone erosions were due to a vaccine, they did emphasize that there was no permanent damage or symptoms. Specifically, I note that no student developed rotator cuff thinning or defects, glenohumeral arthritis, or adhesive capsulitis. If this is what vaccine-specific pathology looks like, it’s quite benign, and it’s not what we see in the vast majority of claims of SIRVA. Dr. Leopold:And one for each of you: Balance out for me the harm that would be caused if your view turns out to be wrong—to individuals and to society more generally—against the benefits that accrue if you’re right. Dr. Srikumaran: Many patients whom I believe have SIRVA convey to me the substantial frustration they have with multiple clinicians who are either unaware of the entity or deny its existence and their experience. By not even acknowledging the possibility of the entity, we are effectively gaslighting a group of people and their experiences, thereby damaging the physician-patient relationship. Further, medicine in general does not have a great history of acknowledging the symptoms and experiences of patients with poorly understood conditions (like women and endometriosis and NFL players and traumatic encephalopathy, for example). I am not aware of any evidence that suggests SIRVA has led to vaccine hesitancy or other harm. If we accept the risk of SIRVA being on the order of 1 in 150,000, that is comparable to the risk of death from bicycling—and here, we are not talking about death, but rather a bursitis that is readily treatable. The potential harm of additional vaccine hesitancy is unlikely by acknowledging this small risk that patients readily accept with so many other interventions. So, when considering the prospect of not believing our patients, I feel the burden of evidence rests on the skeptics. In other words, the theory of SIRVA should be disproven before we suggest to our patients their experience is not valid or is completely “in their heads.” Further, I think the danger is greater in our not acknowledging an entity that has reasonable (considering its rarity) and increasing evidence for its existence and feeds into the notion medicine and the pharmaceutical industry may be hiding potential complications from the general public. I feel this erosion of trust is much more likely to result in future vaccine hesitancy than acknowledging an extremely rare, non-life-threatening, readily treatable condition. If we don’t acknowledge the entity, it is more likely to become a chronic condition that is more difficult to treat requiring further or more invasive treatments. By acknowledging the possibility of SIRVA and treating it early, we can maintain patient trust and avoid the long-term conditions that are compensated by the federal Vaccine Injury Compensation Program (VICP), thereby also reducing the economic burden to the program. Dollars spent for the VICP are likely well spent as the program was specifically created to avoid civil litigation of healthcare practitioners and vaccine manufacturers that could lead to unstable vaccine prices or supply or lack of their development due to the perceived risk of civil litigation. There are other advantages to acknowledging SIRVA, including focusing on prevention and mitigation strategies. In addition to educating practitioners on proper injection techniques to avoid the bursa, we can look at an alternative intramuscular injection site that is not adjacent to a bursa, such as the mid-lateral thigh. Personally, when I get a vaccination, I abduct my shoulder, put two fingers from my opposite arm just below my acromion, and ask the injection to be placed below those two fingers, far away from my bursa. I think we can take a nuanced approach to fully acknowledging the entity while also strongly suggesting SIRVA does not cause things like rotator cuff tears and arthritis. In my opinion the message can be simple. “SIRVA is likely associated with vaccinations at a rare rate of 1 in 150,000 or so. SIRVA does not cause chronic shoulder conditions like rotator cuff tears and arthritis. It is largely preventable, and here are some tips to avoid it. Even if not prevented, in the rare instance it occurs, we can treat it effectively with steroids/injections and rarely an arthroscopic debridement in patients whose symptoms persist.” Dr. Ring: I can’t see much potential harm in being strictly scientific, particularly when it comes to etiology, and particularly when the intervention in question—vaccination—can be considered a public-health miracle [17]. The discovery that peptic ulcer disease was related to infection with H. pylori and not to stress was a matter of intense critical thinking and experimentation [27]. To discover a vaccine-specific pathology, we are going to need to be similarly scientific, and I admire how Hirsiger and colleagues are attempting to meet that threshold [20]. I have found that most people have at least one misconception about their illnesses, and most of those misconceptions seem (at least superficially) consistent with common sense. Given the strong association between unhelpful thinking and pain intensity [7, 8, 28], clinicians who reinforce a person’s misconceptions about pain will likely cause that person more pain. Identifying this type of unhelpful thinking and gently reorienting it is as important as any other step in diagnosis and treatment. It’s important to avoid stigmatizing the human tendency to misinterpret symptoms, as embodied in the phrase “all in your head.” Perception is indeed “all in the head,” and that should lead us to compassion, not marginalization [30]. We can teach people to place the vaccination in a piece of muscle you can grab between your fingers, either on the shoulder or thigh, without telling them that if they don’t, they risk harming the shoulder. We can deliver the instruction without the nocebo. Imagine a society in which all potentially harmful ideas are held in cautious reserve until well-supported by experimental evidence. The potential for improved health and diminished harm from more practiced and automatic critical thinking and debiasing strategies may represent another public health miracle.