Authors’ reply: a commentary on “Assessing kidney outcomes in childhood-onset lupus nephritis: role of National Institutes of Health-modified histological indices”
Article information
To the editor,
We thank the editors for the opportunity to respond and appreciate the interest of Drs. Ma and Zhu [1] in our study of kidney outcomes in childhood-onset lupus nephritis [2].
They note that our Cox model included 6 covariates but only 18 kidney function impairment (KFI) events, resulting in a low events-per-variable ratio [3]. We agree this increases the risk of overfitting and reduces estimate precision. Our intent, however, was etiologic and exploratory—to examine associations between histologic features, treatment, and long-term kidney function in an uncommon pediatric condition—rather than to derive a stand-alone prediction model. We also acknowledged the limitations of sample size and single-center design and highlighted the need for larger, preferably multicenter, cohorts for validation.
Secondly, regarding covariate selection, we included variables associated with KFI in univariate analyses (tubular atrophy and mycophenolate mofetil [MMF] use) and prespecified indicators of disease severity and potential confounders. When the goal is confounder control, restricting covariates to those with significant univariate associations can omit clinically important factors and bias adjusted estimates. Nonetheless, we agree that model complexity must be balanced against sample size [4].
Finally, because the chronicity index includes tubular atrophy and interstitial fibrosis, modelling both the composite and its components would introduce collinearity and may further destabilize estimates. Our a priori focus was to identify which chronic tubulointerstitial lesions were most informative. In a sensitivity analysis, we refit the model by replacing tubular atrophy and interstitial fibrosis with the total chronicity index (per 1-point increase); the association between greater chronic damage and KFI remained in the same direction and of similar magnitude, and our conclusions were unchanged (Table 1).
Primary and sensitivity Cox regression models of kidney function impairment: chronicity components versus total chronicity index
We are grateful for their constructive critique. Within these constraints, our data support that chronic tubulointerstitial damage—particularly tubular atrophy—and MMF-based maintenance therapy are associated with long-term kidney outcomes in childhood-onset lupus nephritis.
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Funding
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.