We greatly appreciate the thoughtful commentary by Dr. Dadashpour and Dr. Golestanieraghi on our article entitled “ Association between vasopressin administration and mortality in patients with cardiogenic shock .” We would like to offer a response to their insightful comments and share their opinion that further prospective research on the use of vasopressin in cardiogenic shock is warranted.
As Dadashpour et al point out, our study is among the largest to examine the association between vasopressin use and outcomes in patients with cardiogenic shock (CS). Despite being substantially more severely ill, patients who received vasopressin had similar unadjusted in-hospital mortality compared to those who did not receive vasopressin implying better adjusted outcomes. Accordingly, we observed significantly lower in-hospital mortality in the vasopressin group following a robust multivariable adjustment with propensity adjustment. Specifically, in the sub-group of patients who required high doses vasopressor therapy, vasopressin was associated with significantly lower in-hospital mortality (unadjusted OR 0.54, 95% CI, 0.32-0.92, P =.02). This is physiologically plausible insofar as patients who require high doses of vasopressors presumably have hyporesponsiveness to catecholamines and previously have been hypothesized to benefit from vasopressin. , These associations led us to conclude that vasopressin may warrant dedicated prospective evaluation in selected patients with CS, particularly patients with refractory shock who may have a mixed cardiogenic-vasodilatory phenotype. Importantly, our retrospective analysis was conducted in a single center and spanned several years of clinical practice. As such, our data was subject to several inherent limitations which we acknowledge in our discussion, most notably the lack of invasive hemodynamic data to confirm the presence of mixed shock, the possibility of selection bias or confounding by indication, and the lack of standardized protocols for initiating or escalating vasoactive medications during the study period.
Dadashpour et al raise appropriate concerns about selection bias and the preferential use of vasopressin in patients displaying features of septic shock. The higher prevalence of sepsis and intravenous antibiotic administration in the vasopressin group supports this idea, but we do not believe that this explains our findings. A sensitivity analysis excluding patients who were diagnosed with sepsis demonstrated lower adjusted in-hospital mortality in the vasopressin group. This suggests that presence of sepsis alone would not completely account for lower adjusted mortality with vasopressin, particularly considering that patients with concomitant sepsis and cardiogenic shock are a known high-risk group expected to have worse outcomes. Many patients with CS have a concomitant inflammatory response including fever, neutrophilia and systemic vasodilation which can mimic sepsis, and such patients (another known high-risk group) could have been misdiagnosed as having sepsis. This would presumably represent a mixed shock phenotype which may benefit from vasopressin, although this is purely a hypothesis-generating statement given the absence of invasive hemodynamic data and the possibility of residual confounding.
Dadashpour and Golestanieraghi make a pertinent point about the use of dopamine in our cohort. Dopamine use has been associated with harm when compared to norepinephrine in patients with CS, and the lower incidence of dopamine use in the vasopressin cohort may account for some of the observed benefit of vasopressin. Indeed, one possible mechanism to explain our results may be reduced exposure to catecholamines and a lower incidence of catecholamine toxicity. This is one of the proposed mechanisms by which vasopressin exerts benefits in other populations, and would be important to demonstrate in patients with CS. Ultimately, further studies of vasopressin which reflect contemporary clinical practice are essential, and ideally would include more granular data regarding real-time hemodynamic changes in response to vasopressin initiation and discontinuation.
In summary, we concur with the astute remarks of Drs Dadashpour and Golestanieraghi regarding the limitations of our analysis, many of which are shared with all retrospective cohort studies and we have attempted to account for analytically to the extent possible. That our main findings were robust to a variety of different adjustment strategies and plausible subgroup analyses argues against our findings being spurious. We have framed our findings as preliminary and hypothesis-generating, and we do not advocate for the routine use of vasopressin in all patients with CS. Instead, our work is intended to advocate for future prospective studies examining the utility of vasopressin in CS, and we hypothesize that there may be a subgroup of CS patients (perhaps those with mixed cardiogenic-vasodilatory shock) in whom it could provide benefit. Indeed, by highlighting the limitations of our study, we hope that future research will be able to overcome similar pitfalls by incorporating comprehensive hemodynamic phenotyping, contemporary MCS and vasoactive medication protocols, and close monitoring for vasopressin-specific cardiovascular adverse effects.
Sincerely
Dhruv Sarma et al
Additional author(s)
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
