Using absolute liver attenuation <48 Hounsfield units (HU), the prevalence was 7% (n = 42/584). Steatosis was reported for only 12 of 42 (28%) of these patients and was documented in clinical notes in only 3 of those cases
232 (40%) had liver enzymes available within 24 hrs of CT scan. 79 had elevated ALT values; steatosis accounted for only 22% of those with elevated ALT values
In those with liver attenuation < 48 HU, nearly all had abnormal ALT values and the median was 52 U/L
Patients with steatosis had an 8-fold likelihood of overweight/obesity
These findings are similar to an adult study of 1290 patients in which 26% had steatosis but only 5% had those findings identified and documented beyond the radiology report. (N Kutaiba et al. J Med Imaging Radiat Oncol 2019; 3: 431-8)
CT findings are considered much more accurate than ultrasonography.
The authors argue that identification of NAFLD is “crucial” to allow for further specialty evaluation and to exclude secondary causes of steatosis.
My take: This study shows that there is an opportunity to improve identification of incidental steatosis. If identified, this can/should be addressed by their primary care team to emphasize improved diet choices and physical activity.
A recent retrospective single-center study (J Satkunasingham et al. Liver Transplantation 2018; 24: 470-77) shows that MRI is a good tool to assess hepatic steatosis. In total there were 144 liver donor candidates; a subset of 32 underwent liver biopsy.
When examining magnetic resonance spectroscopy (MRS) and MRI -proton pump density fat fraction (PDFF), the authors found that MRS-PDFF and MRI-PDFF had 95% and 100% negative predictive value in identifying patients with clinically significant histologic steatosis (≥10%).
The associated editorial by James Trotter (pg 457-58) makes several important points:
Currently living donor transplantation in the U.S. accounts for 4% of all transplants
In his center (and most centers), protocol biopsy are not required prior to liver donation. The main indications for donor liver biopsy are biochemical dysfunction or steatosis on imaging studies.
My take (borrowed from editorial): “Noninvasive estimation of hepatic steatosis is sufficiently accurate to forgo liver biopsy in most donors, although ultimately this decision will continue to rest with the individual center.”
A recent study (JB Schwimmer et al. Hepatology 2015; 1887-95, editorial Vos MB, pages 1779-80) examines the accuracy of magnetic resonance imaging (MRI) compared with liver histology in children with nonalcoholic fatty liver disease.
This prospective validation study enrolled 174 children with a mean age of 14 years. The MRI estimated the liver proton density fat fraction (PDFF).
Liver MRI-PDFF correlated with steatosis grade; the correlation was particularly strong at high and low end values. Thus, a very low MRI-PDFF was highly likely to predict a steatosis grade 0 or 1 while a very high value corresponded to high steatosis levels.
Liver MRI-PDFF was weaker in children with stage 2-4 fibrosis than in children with no fibrosis
The editorial notes that this study “is one of hundreds now published in the literature on MRI and NAFLD…The superiority of MR-based methods…over ultrasound is clear. The question is why are we still ordering abdominal ultrasounds to diagnose NAFLD in children?” The barriers for usage of MRI include cost, potential sedation, and nonuniform methods for MRI usage.
The paper conclude that “MRI is not yet sufficient to replace liver biopsy in children.” The editorial also indicates that the MRI era is fast approaching but not viable today.
Take-home point: Due to the huge numbers of patients with pediatric NAFLD, MRI remains a terrific area for research but remains problematic in clinical practice. Given the expense of MRI, until its use can reduce liver biopsies or improve management, its role is likely to remain limited.
A provocative study (Patel KR, White FV, Deutsh GH. JPGN 2015; 60: 152-58) shows that hepatic steatosis/fatty liver is prevalent at birth in at-risk stillborns.
The authors retrospectively examined autospy results from 33 stillborns (20-40 weeks) delivered to women with diabetes (pregestational or gestational) along with 48 age-matched controls. The majority of women (54%) were African American women; 27% were white and 9% were hispanic.
Hepatic steatosis was common and severe in the stillborns of diabetic women. Prevalence: 78.8% (26/33) compared with 16.6% (8/48) of controls.
No direct correlation was identified between steatosis and glycemic control.
Whether nonalcoholic fatty liver disease (NAFLD) begins at birth is not known and what happens to the fat in newborns with hepatic steatosis is not clear. This study indicates that maternal diabetes may increase the risk of NAFLD.