Interaction between serum chloride increase and baseline chloride level

  • Yanfei Shen
  • Jing Yan
  • Guolong CaiEmail author
  • Matthew F. Barhight
  • John Brinton

Initial correspondence from Drs. Shen, Yan, and Cai

Dear Editor,

Recently, Dr. Barhight et al. reported an association between chloride increase and high mortality in critically ill children [ 1]. Although well designed, their study presents several limitations. First, the number of deaths in the cohort is small (4%, 71/1935). However, 21 variables were included in the logistic regression model (Table 2). Although no consensus has been reached, the rule of thumb is that at least 10 events per variable [ 2] are needed to avoid bias in multivariable regressions. Thus, the robustness of Barhight’s findings needs careful evaluation. Second, multicollinearity issues resulting from highly correlated predictors also play an important role in bias [ 2]. While the correlations in this study (for instance, the correlation between serum sodium and chloride) may be minimized by the fact that the continuous variables were translated into categorized variables, the result of a multicollinearity test...




Compliance with ethical standards

Conflicts of interest

All authors declare that they have no conflict of interest.

Supplementary material

134_2019_5612_MOESM1_ESM.docx (16 kb)
Supplementary file1 (DOCX 15 kb)


  1. 1.
    Barhight MF, Brinton J, Stidham T, Soranno DE, Faubel S, Griffin BR, Goebel J, Mourani PM, Gist KM (2018) Increase in chloride from baseline is independently associated with mortality in critically ill children. Intensive Care Med 44:2183–2191CrossRefGoogle Scholar
  2. 2.
    Courvoisier DS, Combescure C, Agoritsas T, Gayet-Ageron A, Perneger TV (2011) Performance of logistic regression modeling: beyond the number of events per variable, the role of data structure. J Clin Epidemiol 64:993–1000CrossRefGoogle Scholar
  3. 3.
    Johnson AE, Pollard TJ, Shen L, Lehman LW, Feng M, Ghassemi M, Moody B, Szolovits P, Celi LA, Mark RG (2016) MIMIC-III, a freely accessible critical care database. Sci Data. Google Scholar
  4. 4.
    Vittinghoff E, McCulloch CE (2007) Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol 165:710–718CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  • Yanfei Shen
    • 1
  • Jing Yan
    • 1
  • Guolong Cai
    • 1
    Email author
  • Matthew F. Barhight
    • 2
  • John Brinton
    • 3
  1. 1.Department of Intensive CareZhejiang HospitalHangzhouPeople’s Republic of China
  2. 2.Division of Critical Care, Ann & Robert H. Lurie Children’s Hospital of ChicagoNorthwestern University Feinberg School of MedicineChicagoUSA
  3. 3.Department of Biostatistics and InformaticsUniversity of Colorado School of Public HealthAuroraUSA

Personalised recommendations