To the Editor:
Low tidal volume ventilation (LTVV) is the only known intervention with a mortality benefit in all-comers of acute respiratory distress syndrome (ARDS).(1) However, in short statured patients, with low predicted body weight (PBW), delivering LTVV can be challenging, as strict adherence can lead to very low tidal volumes and increased dead space ventilation (VD/VT); conversely, non-adherence can lead to over-distension, with both outcomes associated with higher mortality.(1–3) We hypothesized that short stature (SS) is associated with higher mortality, and that Ventilatory Ratio (VR), a VD/VT surrogate, and/or tidal volume (VT)/PBW mediated this association. Some of the results of these studies have been previously reported in the form of an abstract .(4)
Methods
We analyzed four NHLBI ARDS randomized clinical trials: a combination of EDEN and FACTT formed the discovery cohort (N=1917) and SAILS and ROSE the replication cohort (N=1751)(5–8). All patients received invasive mechanical ventilation (IMV) and enrolled <48h of ARDS onset. Pre-randomization data were used to test for association with 90-day mortality (primary outcome).
To capture the shortest 10% of the population, SS was defined as PBW≤50kg (equivalent heights, males:1.52m; females:1.57m) and normal stature (NS) as PBW>50kg. We used PBW, rather than height, because PBW had lower missingness. Sensitivity analyses were performed with an adjusted SS group accounting for shortest 10% by sex-males with PBW<60kg and females with PBW<45 kg; or excluding patients with PBW <40kg (a total of 50 patients) to mitigate confounding from syndromic pathology.
VR was calculated as previously described.(9) To test the association of SS with 90-day mortality we built logistic regression models adjusted for demographics (age, sex, race), the PaO2:FiO2 (P/F) ratio, APACHE-III scores, and where available, ARDS molecular phenotype (Hyper-/Hypo-inflammatory)(10). Race and ethnicity were collapsed into binary categories due how data were reported in the earliest studies. Variables with >20% missingness were excluded from analyses. For mediation analyses, necessitated by power calculations, all four cohorts were combined. 90-day mortality was the dependent variable, stature group the independent variable, and VT/PBW or VR were the selected mediators. Models were adjusted for sex, race, age and APACHE-III. Analyses were performed on R Version 4.3.1.
Results
The discovery cohort had 177 (9%) and validation cohort 201 (11%) SS participants. Baseline differences between SS and NS participants are presented in Table 1. Briefly, SS participants were predominantly female, and more frequently Hispanic/LatinX. APACHE-III scores were similar between groups.
Table 1:
Baseline Characteristics of Discovery and Validation Cohorts Categorized by Stature Allocation.
| Discovery Cohort | Validation Cohort | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| NS (N=1740) | SS (N=177) | P | NS (N=1550) | SS (N=201) | P | |
|
| ||||||
| Sex | ||||||
| Female | 746 (42.9%) | 171 (96.6%) | <0.001 | 630 (40.6%) | 196 (97.5%) | <0.001 |
| Age | ||||||
| Median [Q1,Q3] | 50.0 [40.0,61.0] | 52.0 [40.0,66.0] | 0.31 | 56.0 [44.0,65.0] | 61.0 [48.0,69.0] | <0.01 |
| Race | ||||||
| Non-White | 492 (28.3%) | 46 (26.0%) | 0.48 | 290 (18.7%) | 36 (17.9%) | 1 |
| White | 1211 (69.6%) | 125 (70.6%) | 1152 (74.3%) | 141 (70.1%) | ||
| Missing | 37 (2.1%) | 6 (3.4%) | 108 (7.0%) | 24 (11.9%) | ||
| Ethnicity * | ||||||
| Hispanic/Latino | 90 (5.2%) | 19 (10.7%) | <0.05 | 159 (10.3%) | 45 (22.4%) | <0.001 |
| Not Hispanic/Latino | 729 (41.9%) | 80 (45.2%) | 1341 (86.5%) | 149 (74.1%) | ||
| Missing | 921 (52.9%) | 78 (44.1%) | 50 (3.2%) | 7 (3.5%) | ||
| APACHE-III | ||||||
| Median [Q1,Q3] | 90.0 [71.0,113] | 93.0 [75.0,110] | 0.42 | 97.0 [78.0,119] | 96.0 [79.0,119] | 0.96 |
| Missing | 52 (3.0%) | 2 (1.1%) | 114 (7.4%) | 18 (9.0%) | ||
| Class Allocation ** | ||||||
| Hyperinflammatory | 233 (13.4%) | 15 (8.5%) | 0.16 | 594 (38.3%) | 81 (40.3%) | 0.64 |
| Hypoinflammatory | 683 (39.3%) | 54 (30.5%) | 956 (61.7%) | 120 (59.7%) | ||
| Missing | 824 (47.4%) | 108 (61.0%) | ||||
| P/F Ratio | ||||||
| Median [Q1,Q3] | 120 [83.0,171] | 125 [76.5,176] | 0.67 | 119 [88.0,150] | 125 [91.0,159] | 0.21 |
| Missing | 126 (7.2%) | 10 (5.6%) | 169 (10.9%) | 20 (10.0%) | ||
| VR | ||||||
| Median [Q1,Q3] | 1.82 [1.45,2.30] | 2.13 [1.72,2.84] | <0.001 | 1.84 [1.47,2.26] | 2.15 [1.77,2.69] | <0.001 |
| Missing | 111 (6.4%) | 8 (4.5%) | 154 (9.9%) | 28 (13.9%) | ||
| Minute Ventilation | ||||||
| Median [Q1,Q3] | 11.4 [9.20,14.1] | 9.60 [7.50,11.5] | <0.001 | 11.0 [9.00,13.3] | 8.95 [7.60,10.6] | <0.001 |
| Missing | 63 (3.6%) | 6 (3.4%) | 100 (6.5%) | 17 (8.5%) | ||
| Tidal Volume (V T ) | ||||||
| Median [Q1,Q3] | 450 [390,500] | 340 [290,400] | <0.001 | 400 [360,450] | 300 [280,350] | <0.001 |
| Missing | 236 (13.6%) | 20 (11.3%) | 292 (18.8%) | 39 (19.4%) | ||
| V T /PBW | ||||||
| Median [Q1,Q3] | 6.62 [6.00,7.81] | 7.61 [6.28,9.13] | <0.001 | 6.06 [5.92,6.68] | 6.93 [6.07,7.92] | <0.001 |
| Missing | 235 (13.5%) | 20 (11.3%) | 292 (18.8%) | 39 (19.4%) | ||
| pH | ||||||
| Median [Q1,Q3] | 7.37 [7.31,7.42] | 7.37 [7.31,7.42] | 0.99 | 7.35 [7.29,7.40] | 7.34 [7.26,7.40] | 0.15 |
| Missing | 53 (3.0%) | 2 (1.1%) | 62 (4.0%) | 12 (6.0%) | ||
| paCO2 | ||||||
| Median [Q1,Q3] | 39.0 [34.0,45.0] | 39.0 [34.0,45.5] | 0.78 | 41.0 [35.0,48.0] | 41.0 [33.0,48.0] | 0.57 |
| Missing | 53 (3.0%) | 2 (1.1%) | 62 (4.0%) | 12 (6.0%) | ||
| Plateau Pressure | ||||||
| Median [Q1,Q3] | 24.0 [20.0,29.0] | 27.0 [21.5,30.0] | <0.05 | 25.0 [21.0,28.0] | 24.0 [21.0,30.0] | 0.39 |
| Missing | 552 (31.7%) | 51 (28.7%) | 687 (44.3%) | 95 (47.3%) | ||
| Driving Pressure | ||||||
| Median [Q1,Q3] | 15.0 [12.0,19.0] | 16.0 [13.0,20.0] | <0.05 | 13.0 [10.0,17.0] | 14.0 [11.0,18.0] | <0.05 |
| Missing | 558 (32.1%) | 52 (29.2%) | 696 (44.9%) | 97 (48.3%) | ||
| Respiratory Rate | ||||||
| Median [Q1,Q3] | 25.0 [20.0,30.0] | 26.0 [20.0,32.0] | 0.05 | 26.0 [21.0,30.0] | 27.0 [22.0,32.5] | <0.05 |
| Missing | 10 (0.6%) | 0 (0%) | 43 (2.8%) | 2 (1.0%) | ||
Ethnicity was not available for FACTT
Class Allocation was not available for EDEN.
Among respiratory variables, SS and NS participants had similar pH, P/F and PaCO2 in both cohorts (Table 1). In the discovery cohort, SS participants had lower minute ventilation but higher VT/PBW, driving pressure (DP) and VR compared to NS. Similarly, SS participants had lower minute ventilation but higher VT/PBW, DP, VR and respiratory rate in the replication cohort.
SS is associated with 90-day mortality.
SS had higher mortality than NS patients (Discovery: 37% vs. 25%, p<0.001, Replication: 44% vs. 35%, p<0.05); and was significantly associated with mortality in multivariable analysis (Discovery: aOR:2.23, CI:1.48–3.33; Replication: aOR:1.66, CI:1.08–2.52).
VR but not VT/PBW mediates the association between SS and mortality
For mediation analyses, we found SS patients had significantly higher VT/PBW (7.25, IQR:6.15–8.37 vs. 6.22 IQR:5.97–7.27, p<0.001) and VR (2.14, IQR:1.74–2.74 vs 1.82 IQR:1.46–2.28, p<0.001) compared to NS, with SS associated with higher odds of mortality in a multivariable model (Figure 1a). Evaluating the effect of VR on SS and mortality in adjusted mediation, we observed significant total, direct (and indirect effects of VR on the model. Approximately 11% (CI:4–22%, p<0.001, n=3075, 1000 simulations) of the association between SS and mortality was explained/mediated by higher VR, suggesting that VR partially mediates the relationship between SS and mortality (Figure 1b).
Figure 1a: SS is associated with mortality in adjusted multivariable analyses of a combined, four study cohort.
Adjusted multivariable logistic regression was used to test the association of stature categories (SS v NS) and mortality in the combined, four study cohort. Allocation to the Short Stature group was associated with increased 90-day Mortality, with OR 1.94, CI 1.45–2.59, p<0.001 when adjusted for race, age, sex, baseline P/F and APACHE-III; n=3047.
1b, Upper Panel-Ventilatory Ratio (VR) Partially Mediates the Association between Short Stature (SS) and Mortality. The total effect of the model was significant. There was a significant direct effect , and a significant indirect effect of the association between SS and mortality through VR. The average proportion of the association between SS and mortality that was mediated by VR in this model was 0.11 CI 0.04–0.22, p<0.001. Adjusted for race, age, sex, APACHE-III, based on differences observed in Table 1 and the known association between APACHE-III and mortality; n=3075. 1b, Lower Panel-Tidal Volume per Predicted Body Weight (VT/PBW) Does Not Mediate the Association between Short Stature and Mortality. The total effect of the model was found to be significant. It was found that there was a statistically significant direct effect between SS and mortality. However, a statistically significant indirect effect of the association between SS and mortality through VT/PBW was not observed. Adjusted for race, age, sex, APACHE-III ; n=2773. Mediation with non-parametric bootstrapping was performed utilizing the mediation package in R, version 4.5.0. The default quasi-Bayesian approximation was selected for reporting 95% confidence intervals. 1,000 simulations were performed for estimating each model within the mediation analysis. ACME, Average Causal Mediation Effect.
Evaluating the mediating effect of VT/PBW on SS and mortality, we observed a significant total and direct effects in an adjusted model. However, the indirect effect was not significant, suggesting VT/PBW did not mediate the relationship between SS and mortality (Figure 1b).
Sensitivity Analyses of Alternate Short Stature Definitions
As the primary SS cohort was predominantly female, we performed sensitivity analyses to include the lowest 10% of male and of female participants. Another sensitivity analysis excluded participants with PBW<40kg to mitigate likelihood of syndromic diagnosis-related adverse outcomes. In both these sensitivity analyses, the findings of the primary analyses were recapitulated with significant association of SS with mortality and VR mediated this association, but not VT/PBW.
Discussion
Our analyses using data from two decades of post-ARMA research show that short stature is associated with increased mortality in mechanically ventilated ARDS patients. Although we cannot determine mechanism, we did observe that in SS, VR was higher and significantly mediated the association with mortality.
Our findings identify SS patients as a high-risk group and highlight the challenges of using a “one-size-fits-all” approach to ventilatory strategies in ARDS. The fixed calculation of PBW across the entire range of height risks misestimating actual lung volume, particularly at the extremes. In SS, this adherence to fixed volumes risk hypoventilation and increased dead space/VR due to low VT underestimating true lung volume. Our findings of increased VR in SS being a significant mediator of mortality substantiates this hypothesis.
Conversely, intermediate tidal volumes (8–10 mL/Kg) risk increased barotrauma. While PBW/VT did not mediate the increased risk of mortality in SS, in part this may be due to the broader limitation of this variable in the absence of its integration with measures of compliance, whereby intermediate tidal volumes (8–10 mL/kg) may be safe on the proviso that plateau and/or driving pressures are within acceptable limits (30/15 cmH2O, respectively) (11). Unfortunately, we were unable to test the complex interactions of set VT, corresponding DP, and VR due to the high missingness of DP in all cohorts.
Clinically, a hypothetical framework to optimize ventilatory settings in SS could be to target intermediate lung protective ventilation that is DP limited. In the setting of high DP, SS patients might be a subgroup triaged to early ECMO due to the aforementioned challenges of delivering LTVV. However, these hypotheses need prospective testing before implementation. The ongoing PRACTICAL (https://practicalplatform.org/) and PREVENT VILI (https://preventvili.org) RCTs may demonstrate how precision-based ventilation can be optimally implemented; however, generating prospective evidence for best ventilatory practices in SS patients may be additionally challenging, because they represent only 10% of trial populations by our definition. Given that the SS group was older, with more Hispanic/Latinx and female participants, who may have unmeasured confounding from early life exposures and decreased physiologic reserve, our study highlights the importance of recruiting diverse populations to future RCTs and the limited generalizability of currently completed ARDS RCTs.
In conclusion, our study identified SS patients as a high-risk group of ARDS. Clinicians caring for these patients should be aware of their added risk and the potential of VR as a prognostic tool. Future prospective studies need to evaluate personalized ventilation in SS patients to mitigate ventilator-induced lung injury while minimizing VD/VT.
Funding:
R35GM142992, R01HL173531 (PS), K23HL165149 (ACM), R35HL177135 (CSC).
Footnotes
Artificial Intelligence Disclaimer: No artificial intelligence tools were used in writing this manuscript.
References
- 1.Brower RG; Matthay MA; Morris A; Schoenfeld D; Thompson BT; Wheeler ATAN. Ventilation with Lower Tidal Volumes as Compared with Traditional Tidal Volumes for Acute Lung Injury and the Acute Respiratory Distress Syndrome. Surv Anesthesiol 342: 1301–1308, 2000. doi: 10.1097/00132586-200102000-00017. [DOI] [PubMed] [Google Scholar]
- 2.Nuckton TJ, Alonso JA, Kallet RH, Daniel BM, Pittet J-F, Eisner MD, Matthay MA. Pulmonary Dead-Space Fraction as a Risk Factor for Death in the Acute Respiratory Distress Syndrome. N Engl J Med 346: 1281–1286, 2002. doi: 10.1056/NEJMoa012835. [DOI] [PubMed] [Google Scholar]
- 3.Ranieri VM, Suter PM, Tortorella C, De Tullio R, Dayer JM, Brienza A, Bruno F, Slutsky AS. Effect of Mechanical Ventilation on Inflammatory Mediators in Patients With Acute Respiratory Distress Syndrome. JAMA 282: 54, 1999. doi: 10.1001/jama.282.1.54. [DOI] [PubMed] [Google Scholar]
- 4.Monteiro ACC, Bartek B, Calfee CS, Matthay MA, Sinha P. Short Stature Is a Risk Factor for Increased Ventilatory Inefficiency and Is Associated With Mortality in Patients With Acute Respiratory Distress Syndrome in NHLBI Network Trials. In: D14. ACUTE RESPIRATORY FAILURE AND MECHANICAL VENTILATION: EXPANDING OUR BREADTH OF KNOWLEDGE. American Thoracic Society, p. A6917–A6917. [Google Scholar]
- 5.Rice TW, Wheeler AP, Thompson BT, Steingrub J, Hite RD, Moss M, Morris A, Dong N, Rock P. Initial Trophic vs Full Enteral Feeding in Patients With Acute Lung Injury: The EDEN Randomized Trial. JAMA J Am Med Assoc 307: 795–803, 2012. doi: 10.1001/jama.2012.137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.The National Heart L and BIPCTN. Early Neuromuscular Blockade in the Acute Respiratory Distress Syndrome. N Engl J Med 380: 1997–2008, 2019. doi: 10.1056/NEJMoa1901686. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.The National Heart, Lung and BIARDS (ARDS) CTN. Comparison of Two Fluid-Management Strategies in Acute Lung Injury. Wiederman HP; Wheel AP; Bernard GR; Thompson BT; Hayden D; deBoisblanc B; Connors AF; Hite RD; Harabin AL; NHLBI ARDS Clin Trial Network 354: 2564–2575, 2006. doi: 10.1056/NEJMoa062200. [DOI] [PubMed] [Google Scholar]
- 8.The National Heart L and BIACTN. Rosuvastatin for Sepsis-Associated Acute Respiratory Distress Syndrome. N Engl J Med 370: 2191–2200, 2014. doi: 10.1056/NEJMoa1401520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sinha P, Fauvel NJ, Singh S, Soni N. Ventilatory ratio: a simple bedside measure of ventilation. Br J Anaesth 102: 692–7, 2009. doi: 10.1093/bja/aep054. [DOI] [PubMed] [Google Scholar]
- 10.Calfee CS, Delucchi K, Parsons PE, Thompson BT, Ware LB, Matthay MA. Subphenotypes in acute respiratory distress syndrome: Latent class analysis of data from two randomised controlled trials. Lancet Respir Med 2: 611–620, 2014. doi: 10.1016/S2213-2600(14)70097-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Amato MBP, Meade MO, Slutsky AS, Brochard L, Costa ELV, Schoenfeld DA, Stewart TE, Briel M, Talmor D, Mercat A, Richard J-CM, Carvalho CRR, Brower RG. Driving Pressure and Survival in the Acute Respiratory Distress Syndrome. N Engl J Med 372: 747–755, 2015. doi: 10.1056/NEJMsa1410639. [DOI] [PubMed] [Google Scholar]