Fasting intact insulin by mass spectrometry is associated with metabolic dysfunction–associated steatotic liver disease in youth

Abstract

Background:

Fasting intact insulin concentrations can predict metabolic dysfunction–associated steatotic liver disease (MASLD) in adults without diabetes; however, research in youth is limited. We sought to determine whether fasting intact insulin, measured by liquid chromatography-tandem mass spectrometry, is associated with MASLD in children.

Methods:

This retrospective cross-sectional analysis used data and samples from children who participated in studies across 3 universities between 2014 and 2022. Key measurements included fasting intact insulin, ALT, and hepatic steatosis assessed by MRI techniques. MASLD was defined as hepatic steatosis ≥5% by MRI with at least 1 cardiometabolic risk factor. The optimal cutoff points to identify MASLD were determined by maximizing the Youden index, and the AUROC curves were compared using the DeLong test.

Results:

The analysis included 184 children (28% male; 14.9 ± 2.6 y; 57% Hispanic race/ethnicity; body mass index 32.5 ± 8.1 kg/m²; 64% with MASLD, 43% with polycystic ovary syndrome, and 5% with other liver diseases). Fasting intact insulin and ALT levels were significantly higher in children with MASLD (p < 0.05). Fasting intact insulin was strongly associated with MASLD with an AUROC of 0.83 (0.77–0.90), sensitivity of 71%, and specificity of 85%. When combined with ALT (intact insulin × ALT [μU/mL × U/L]), the AUROC was 0.88 (0.83–0.94), with a sensitivity of 89% and specificity of 81%. The improvement in AUROC over intact insulin alone was not statistically significant (p = 0.089) but was statistically significant from ALT (p = 0.022). Optimal cutoff points for intact insulin and intact insulin × ALT were 20 μU/mL and 522 μU/mL × U/L, respectively.

Conclusions:

In pediatric patients, measurements of fasting intact insulin alone and combined with ALT provide a noninvasive strategy for identifying the presence of MASLD.

INTRODUCTION

Metabolic dysfunction–associated steatotic liver disease (MASLD), formerly known as NAFLD, is the most common liver disease in children and is a leading cause of liver-related morbidity and mortality.^1,2 The prevalence of MASLD, once considered rare, has significantly increased and is now estimated to be 16.5% among all US adolescents,³ including 26% among those with obesity.⁴ MASLD is characterized by the presence of hepatic steatosis ≥5%, along with at least 1 out of 5 specified cardiometabolic risk factors, in the absence of heavy alcohol use and other chronic liver diseases. In youth, these cardiometabolic risk factors include increased body mass index (BMI), fasting glucose, blood pressure, triglycerides (TGs), and low HDL-cholesterol.

The current and future health burden of MASLD is tremendous both to the individual and from a public health standpoint. There are no FDA-approved medications or supplements for treating pediatric MASLD, and lifestyle modifications have shown limited effectiveness in reducing hepatic steatosis and serum ALT levels.^5–9 A recent longitudinal study of 51 children (47% diagnosed with MASLD) showed that over a 10-year follow-up, the progression of steatosis was associated with worsening of metabolic disturbances such as elevated ALT and TGs, and 6% developed advanced fibrosis.¹⁰ Another study on children diagnosed with MASLD demonstrated that over 4–11 years, 6% developed type 2 diabetes (T2D),¹¹ which is greater than the expected rate,¹² 3% underwent liver transplants for end-stage liver disease, and 3% died, which is a significant increase over the expected pediatric population mortality rate.¹¹

The gold standard for MASLD diagnosis is histopathologic assessment, obtained by liver biopsy¹³ in addition to cardiometabolic risk factors.¹ The invasive nature of liver biopsy, along with its limitations, including small sample sizes, inter-rater variability, and procedural risks, undermines its reliability as a gold standard.¹⁴ Another diagnostic approach is ultrasound, which lacks sensitivity for accurately measuring hepatic steatosis.¹⁵ The accuracy of FibroScan is problematic in youth with significant obesity, secondary to limitations of the depth of the penetration of the vibration wave.¹⁶ Current recommendations suggest screening overweight children for MASLD at age 10 and older by measuring ALT concentrations, which also lack sensitivity and specificity.¹⁷ MRI is a rapidly advancing method for MASLD prediction that is highly sensitive and specific in detecting hepatic steatosis.^18,19 However, MRI is expensive and less available than blood tests and does not characterize the phenotype of the disease (steatosis vs. steatohepatitis) or predict progression. The lack of an accurate, low-risk, noninvasive screening biomarker has hindered the field of pediatric MASLD, including the slowing of drug development. These considerations have led to an international interest in the development of better noninvasive methods for screening and diagnosis.

There is a strong association between MASLD and insulin resistance (IR)²⁰; however, traditional measurements of IR have underperformed as markers of MASLD.²¹ More specifically, immunoassays frequently used for insulin measurement in previous studies are highly variable, lack standardization, and measure multiple insulin intermediates.²² This challenge was addressed by the development of a multiplexed liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay for intact insulin and C-peptide.²³ A recent study conducted by Bril et al²⁴ evaluated the ability of fasting intact insulin measured by LC-MS/MS to predict MASLD in adults and compared the results to MRI for hepatic steatosis measurement, a 2-hour oral glucose tolerance test, and liver biopsy results. They demonstrated that intact fasting insulin was better than most other clinical variables in predicting MASLD in adults without T2D. Furthermore, when combined with ALT, the combined biomarker (intact insulin × ALT) detected MASLD with an AUROC of 0.94 (0.89–0.99). Intact insulin has not been tested in children and adolescents with or without MASLD.

The main objective of this study was to determine whether intact insulin, measured by a recently validated, high-throughput, multiplexed LC-MS/MS method, is associated with the presence of MASLD in youth. We hypothesized that precise measurement of intact insulin concentrations would enable accurate classification of MASLD in children and adolescents.

METHODS

Study design and population

In this cross-sectional cohort study, we analyzed baseline data from 184 patients aged 7–19 years enrolled in 1 of 8 clinical research studies conducted at Emory University (NCT02461212, NCT03042767, and IRB00088900), University of California San Diego (NCT02513121), and the University of Colorado Anschutz Medical Campus (NCT03041129, NCT02157974, NCT03717935, and NCT03919929). The 8 studies focused on children and adolescents at high risk for or with a known diagnosis of MASLD or polycystic ovary syndrome (PCOS), as well as healthy controls. Inclusion criteria included the ability to undergo MRI, while exclusion criteria encompassed a history of significant alcohol intake, chronic medication use known to induce steatosis or steatohepatitis, a history of bariatric surgery, liver transplant, diabetes, renal disease, and pregnancy. Participants were included in our present analysis if they had the following data available: hepatic steatosis assessed by magnetic resonance imaging-proton density fat fraction (MRI-PDFF) or proton magnetic resonance spectroscopy (¹H-MRS), fasting intact insulin levels, and ALT measurements. Since children with other chronic liver diseases present similarly to those with MASLD, we did not exclude them from our primary study to more accurately reflect the real-world clinical setting (see Supplemental Table S1, http://links.lww.com/HC9/B85, for additional information on the 8 parent studies).

All participants provided written informed consent and assent for their samples and data to be used for future analysis. The research was conducted in accordance with both the Declarations of Helsinki and Istanbul, and the study protocol was approved by the institutional review boards at Emory University, Children’s Healthcare of Atlanta (Atlanta, GA), the University of California San Diego (San Diego, CA), and the University of Colorado Anschutz (Aurora, CO).

MASLD status

MASLD was defined as hepatic fat ≥5% measured using either MRI-PDFF or ¹H-MRS, along with at least 1 marker of cardiometabolic dysfunction.¹ In youth, these cardiometabolic risk factors include BMI ≥85th percentile for age/sex (BMI z-score ≥+1), fasting serum glucose ≥100 mg/dL, blood pressure ≥130/80 mm Hg, plasma TGs ≥100 mg/dL (age <10 y) or ≥150 mg/dL (age ≥10 y), and plasma high-density lipoprotein-cholesterol (HDL-c) ≤40 mg/dL.¹

Biochemical data

Blood samples from the baseline visit of all studies were drawn after an overnight fast (minimum of 8 h), processed, and stored in a freezer at −80°C. Clinical and laboratory assessments of the combined cohort were similar in each parent study and are summarized in Table 1. Anthropometric measurements, including height (cm) and weight (kg), were obtained using standard operating procedures. BMI was calculated as weight in kilograms divided by the square of height in meters squared. Demographic information on age, sex, and race/ethnicity was collected using questionnaires or medical record reviews. Laboratory tests performed in all parent studies included transaminases (ALT and AST), fasting glucose, TGs, LDL-cholesterol, and HDL-cholesterol. Fasting glucose and intact insulin were used to calculate the homeostasis model assessment of insulin resistance (HOMA-IR).²⁵ HOMA-IR was calculated as fasting glucose (mg/dL) × fasting insulin (µU/mL)/405.

TABLE 1.

Sociodemographic and clinical characteristics of the sample stratified by MASLD status

Parameter	Non-MASLD, N = 67	MASLD, N = 117	p
Age, y	15 (2.78)	15 (2.49)	0.400
Male (n, %)	12 (18)	40 (34)	0.026
Hispanic (n, %)	22 (33)	83 (71)	<0.001
% Hepatic steatosis	2.82 (2.19–3.62)	13.96 (8.90–21.00)	<0.001
BMI (kg/m2)	28.5 (8.9)	34.9 (6.6)	<0.001
BMI %ile	96.2 (73.3–98.4)	98.7 (97.8–99.4)	<0.001
ALT (IU/L)	21.5 (17.0–29.0)	48.0 (34.0–81.0)	<0.001
AST (IU/L)	27.0 (19.0–36.4)	36.0 (29.0–50.0)	<0.001
Fasting glucose (mg/dL)	88.0 (83.0–93.0)	92.0 (86.0–98.0)	0.003
TG (mg/dL)	95.0 (63.0–123.0)	115.0 (87.0–181.0)	<0.001
LDL-c (mg/dL)	77.4 (66.0–94.0)	83.0 (69.8–99.8)	0.130
HDL-c (mg/dL)	45.0 (36.0–58.0)	34.0 (29.0–40.0)	<0.001
Intact insulin (µU/mL)	12.0 (7.0–18.0)	27.0 (18.0–38.0)	<0.001
C-peptide (ng/mL)	1.8 (1.1–2.4)	3.5 (2.5–4.5)	<0.001
IR score	50.0 (14.0–81.0)	99.0 (86.0–100.0)	<0.001
HOMA-IR	2.6 (1.3–4.1)	6.0 (3.8–8.3)	<0.001

Note: The statistics shown are mean (SD), median (IQR), or n (%). Statistical significance was considered as p < 0.05, indicated in bold.

Abbreviations: BMI, body mass index; c, cholesterol; HOMA-IR, homeostasis model assessment of insulin resistance; IR, insulin resistance; MASLD, metabolic dysfunction–associated steatotic liver disease; TG, triglycerides.

Intact insulin and C-peptide measurements

Quest Diagnostics Nichols Institute performed measurements of intact insulin and C-peptide as described.^23,26 Briefly, stored patient serum samples were thawed, vortexed, and transferred to the deck of a Hamilton Microlab Star robotic liquid handler. Next, the samples were plated and mixed with the addition of internal standards (bovine insulin and a stable isotopically labeled [13C/15N] C-peptide) and Cleanascite delipidation reagent. Samples were centrifuged, and the supernatant was plated with a suspension of magnetic beads prepared for immunocapture. After 1 hour of incubation, serum supernatant removal, bead washing, and extraction were performed. During extraction, Trizma base was added to the wells of the elution plate to enhance peptide stability before injection into the LC system. Analytical separation using LC-MS/MS was performed using a fully automated online 2-dimensional LC system (TurboFlow Aria TLX-4, Thermo-Fisher). Tandem mass spectrometry data were acquired using a 6490 Triple Quadrupole mass spectrometer with selected reaction monitoring in positive ion mode. Intact insulin and C-peptide have demonstrated stability for a minimum of 39 weeks when stored at −80°C and over at least 5 freeze/thaw cycles.²³ The IR score was calculated using intact insulin and C-peptide values as described.²⁷

Statistical analysis

All variables were evaluated for normality. The sociodemographic and clinical characteristics of the study participants are summarized in Table 1 using counts and percentages for categorical variables and mean and SD or median and IQR for continuous variables, depending on the distribution. Using the R software program version 4.2.3, a small proportion of missing data (11%) were imputed using the R package mice with the predictive mean matching method.²⁸ Differences in sample characteristics between MASLD and non-MASLD controls were assessed using the Mann-Whitney U test (Wilcoxon rank-sum test) for continuous variables and the Fisher exact test for categorical variables. The cutoff points for determining test positive or test negative status were selected based on the Youden index. Sensitivity, specificity, positive predictive values, and negative predictive values for intact insulin, ALT, C-peptide, and the IR score were calculated using MRI-PDFF/¹H-MRS data in combination with the presence of at least 1 of the 5 cardiometabolic risk factors for the diagnosis of MASLD. Receiver operating characteristic curves were plotted, and the AUROC with its 95% CI was calculated to represent the association of the biomarkers with MASLD versus non-MASLD. Comparisons between AUROCs were performed by applying the receiver operating characteristic test of DeLong using the roc.test function of the pROC library in R.²⁹ Statistical significance was set at p < 0.05. In a separate sensitivity analysis, children with other liver diseases were excluded to evaluate the association of fasting intact insulin with MASLD in a more homogenous cohort.

RESULTS

Patient characteristics

Our combined cross-sectional cohort included data from 184 participants. A total of 117 participants met the diagnostic criteria for MASLD based on the presence of hepatic steatosis ≥5% coupled with at least 1 of 5 cardiometabolic criteria, as defined above.¹ Our control group consisted of 67 subjects, including 10 patients with other liver diseases such as primary sclerosing cholangitis, alpha-1 antitrypsin deficiency, and autoimmune hepatitis. Our cohort also included 80 females with polycystic ovary syndrome (PCOS: 76% with MASLD and 24% without MASLD).

Clinical and sociodemographic characteristics of the 184 participants are described in Table 1. Compared to the non-MASLD group, the MASLD group had a higher percentage of male subjects (34% vs. 18%, respectively), Hispanics (71% vs. 33%), and obese individuals (74% vs. 43%) (all p < 0.05). The MASLD group also had significantly higher fasting glucose, ALT, AST, HOMA-IR, TGs, intact insulin, intact C-peptide, and IR scores with lower HDL-c levels than the non-MASLD group (all p < 0.05).

Use of intact insulin for the diagnosis of MASLD in children

Fasting intact insulin was strongly associated with the presence of MASLD in children and adolescents, with an AUROC of 0.83 (0.77–0.90), sensitivity of 71%, and specificity of 85% at an optimal cutoff of 20 μU/mL. ALT detected the presence of MASLD with relatively high accuracy in children with an AUROC of 0.82 (0.75–0.89) with a sensitivity of 83% and specificity of 76% (Figure 1). We then explored whether ALT × fasting plasma intact insulin provided a stronger association with MASLD than each of these tests alone. Intact insulin (μU/mL) × ALT (U/L) identified MASLD with an AUROC of 0.88 (0.83–0.94). The improvement in AUROC compared with ALT alone was statistically significant (p = 0.022; Figure 1), and there was a trend toward better performance compared with intact insulin alone (p = 0.089). When using the combined biomarker and an optimal cutoff point of 522 µU/mL × U/L, the sensitivity, specificity, positive predictive value, and negative predictive value were 89%, 81%, 89%, and 81%, respectively. When we excluded children with other liver diseases (n = 10), intact insulin performed similarly with an AUROC of 0.82, while the AUROC for ALT increased to 0.90 (Figure 2). In addition, the combination of intact insulin × ALT demonstrated high accuracy in its association with MASLD, achieving an AUROC of 0.92 (0.87–0.97), sensitivity of 91%, and specificity of 82%.

FIGURE 1.

ROC curves demonstrating the association of MASLD with intact insulin, ALT, or intact insulin × ALT (n = 184). p values represent the comparison against intact insulin × ALT. p values <0.05 were considered statistically significant. Abbreviations: MASLD, metabolic dysfunction–associated steatotic liver disease; ROC, receiver operating characteristic.

FIGURE 2.

ROC curves demonstrating the association of MASLD with intact insulin, ALT, or intact insulin × ALT excluding children with other liver diseases (n = 174). p values represent the comparison against intact insulin × ALT. p values <0.05 were considered statistically significant. Abbreviations: MASLD, metabolic dysfunction–associated steatotic liver disease; ROC, receiver operating characteristic.

Comparing intact insulin with other MASLD biomarkers

Based on the findings in adults without T2D,²⁴ we also evaluated how the product of intact insulin and ALT performed compared to intact C-peptide and the IR score for diagnosing MASLD in children. C-peptide identified MASLD with an AUROC curve of 0.86 (0.80–0.91), a sensitivity of 77%, and a specificity of 82%. Similarly, the IR score, based on fasting intact insulin and C-peptide, had an AUROC curve of 0.85 (0.79–0.91) with a sensitivity of 75% and specificity of 84% (Supplemental Figure S1, http://links.lww.com/HC9/B85). Both biomarkers had lower AUROC values than intact insulin × ALT; however, differences did not reach statistical significance.

DISCUSSION

The rising prevalence of pediatric MASLD underscores the need for noninvasive and accurate screening methods.³⁰ In this study, we found that fasting intact insulin measured using LC-MS/MS is strongly associated with MASLD in children and adolescents. Furthermore, when combining this measurement with plasma ALT levels, we were able to identify the presence of MASLD detected by MRI techniques with positive and negative predictive values >80%. These results are promising, potentially offering a feasible option for diagnosing pediatric MASLD in both clinical practice and large epidemiological studies.

Traditional measures of IR versus intact insulin for identifying MASLD

IR contributes to MASLD through multiple mechanisms, including the upregulation of de novo lipogenesis by hyperinsulinemia and increased free fatty acid flux to the liver through decreased inhibition of lipolysis.³¹ In turn, hepatic steatosis may also promote IR independent of visceral fat and intramyocellular lipid content.³² Intrahepatic lipid accumulation is linked to reduced insulin clearance and hepatic IR in obese youth, irrespective of ethnic background.³³ Conversely, the reduction of hepatic steatosis is associated with improvements in insulin sensitivity.^34,35 It is unclear whether hepatic steatosis is a cause or a consequence of alterations in insulin sensitivity. Nevertheless, it is evident that there is a strong relationship between MASLD and IR.³⁶

Traditional measures of IR such as HOMA-IR, triglyceride-glucose (TyG) index, triglyceride/high-density lipoprotein-cholesterol ratio (TG/HDL-c), and the metabolic score for insulin resistance have not been fully successful in predicting MASLD.²¹ For example, a study in adults showed that HOMA-IR predicted MASLD with an AUROC of 0.78, a sensitivity of 72%, and a specificity of 73%. This performance is slightly lower than the AUROC of 0.88 and sensitivity of 89% for intact insulin × ALT observed in our analysis. Possible reasons for this discrepancy may include the lack of standardization of immunoassays that measure not only insulin but also proinsulin and other intermediates.²² To address these limitations, this study used state-of-the-art LC-MS/MS methods that specifically quantify only metabolically active intact insulin molecules.²³

Investigations of the usefulness of traditional IR markers for detecting MASLD in children and adolescents are limited.³⁷ One population-based study reported that modified TyG indices, which combine obesity-related parameters such as BMI and waist circumference with the TyG index, were superior to the TyG index alone for predicting pediatric MASLD.³⁸ However, the highest-performing parameter, the TyG-waist-to-height ratio, had only an AUROC of 0.78 with a sensitivity of 61% and specificity of 83%. Another study found that fasting insulin was the best predictor of MASLD in 108 obese children at a cutoff point of 18.9 μU/mL with an AUROC of 0.83, sensitivity of 75%, and specificity of 87% compared to the oral glucose tolerance test and HOMA-IR.³⁹ Interestingly, this cutoff point was similar to the optimal cutoff of 20 μU/mL for fasting intact insulin observed in our analysis. Moreover, in a study of 76 overweight/obese youth, the quantitative insulin-sensitivity check index and HOMA-IR demonstrated similar AUROC curve values of 0.81 and 0.82, respectively, in differentiating MASLD from normal liver histology; however, quantitative insulin-sensitivity check index had a higher sensitivity of 80%.⁴⁰ In this study, we demonstrated that fasting intact insulin combined with ALT outperformed these metrics, as well as the IR score and C-peptide, with a higher AUROC curve value and sensitivity for detecting MASLD. These findings highlight the added value of implementing LC-MS/MS to precisely measure intact insulin molecules.

Comparison of intact insulin for detecting MASLD in adults versus children and adolescents

In adults without T2D, fasting intact insulin, also measured using LC-MS/MS, predicted MASLD with high accuracy (AUROC 0.90 [0.84–0.96], sensitivity 93%, specificity 71%) and was better than clinical variables including insulin measured by radioimmunoassay, AST, TG, HDL-c, glucose, HbA1c, and BMI.²⁴ This AUROC for fasting intact insulin in adults was higher than the AUROC of 0.83 observed in our analysis of children; however, our specificity was 14% higher. In addition, the optimal cutoff point of 10.5 μU/mL in adults was almost 50% lower than the cutoff of 20 μU/mL that we observed in children. This may be due to the influence of puberty since hormonal changes that occur during this time can alter insulin sensitivity, leading to higher insulin concentrations in children and adolescents.⁴¹

In adults, intact insulin × ALT was significantly better than previously validated noninvasive scores such as the TyG index, NAFLD – liver fat score (NAFLD-LFS), and hepatic steatosis index. The AUROC for intact insulin × ALT in adults was 0.06 greater than that observed in the current study among children, achieving a value of 0.94. Again, the optimal cutoff of 290 μU/mL × U/L in adults was almost 50% lower than the cutoff of 522 μU/mL × U/L in children, likely due to the effects of IR during puberty.

Children with other liver diseases, such as autoimmune hepatitis and primary sclerosing cholangitis, often present with abnormally elevated ALT levels.^42,43 Therefore, when we excluded the 10 children with other liver diseases from our analysis, the AUROC curve value for ALT increased to 0.90. This increase raised the AUROC of intact insulin × ALT to 0.92, similar to the value observed in adults.

Comparisons with other current biomarkers for pediatric MASLD

In the clinical setting, hepatic transaminases, such as ALT, are commonly used for detecting MASLD owing to their cost and availability.^15,44 However, recent studies have suggested that they are not sensitive enough to identify adolescents with MASLD.¹⁷ In a cross-sectional study of 121 children, ALT detected MASLD with medium accuracy (AUROC 0.74, sensitivity 79%, and specificity 50%) using the upper limit of normal as the cutoff (22 IU/L for girls and 26 IU/L for boys) and was comparable to ultrasound.¹⁷ Applying a threshold of ALT ≥40 IU/L improved the specificity, but the sensitivity decreased to <45%. In our study, the cutoff for ALT that maximized the Youden index of 30 U/L was close to the upper limit of normal for youth and showed relatively high accuracy with an AUROC of 0.82, a sensitivity of 83%, and a specificity of 76%. Although ALT alone was similar in accuracy to fasting intact insulin alone, the combination of biomarkers (intact insulin × ALT), which together reflect IR and liver injury/inflammation, provides a highly accurate measure for detecting MASLD in youth.

In recent years, several other noninvasive circulating biomarkers have been investigated for their potential in detecting pediatric MASLD.^45,46 These include interleukins, adiponectin, FGF21, chemerin, and cytokeratin-18, among others. Most of these biomarkers predicted MASLD with AUROC values >0.70; however, most are not liver-specific factors. Thus, their role as diagnostic markers of MASLD remains questionable.⁴⁵

Other studies have demonstrated that the combination of multiple biomarkers and/or the addition of clinical variables can increase the diagnostic accuracy of individual biomarkers.⁴⁶ For example, Khusial and colleagues developed a screening panel for pediatric MASLD using a machine-learning approach that included high-resolution metabolomics and clinical phenotype data. The panel consisted of 11 metabolite features, including the amino acids serine, leucine/isoleucine, and tryptophan, as well as waist circumference, whole-body insulin sensitivity index, and TGs. This screening panel predicted MASLD with an AUROC of 0.92, a sensitivity of 73%, and a specificity of 94%.⁴⁷ Moreover, a modified version of this panel was validated in an external cohort and performed with similar accuracy.⁴⁸ Therefore, LC-MS/MS-based metabolomics, in combination with clinical phenotype data, including measures of IR, may offer additional accuracy in diagnosing MASLD.

In another study, Carreau et al⁴⁹ developed a clinical prediction score, termed the PCOS-HS index, to identify MASLD in obese girls with PCOS, a population at high risk for MASLD. The index was based on BMI percentile, waist circumference, ALT, and sex hormone binding globulin and had an AUROC of 0.81, a sensitivity of 82%, and a specificity of 69% for predicting hepatic steatosis. Although intact insulin × ALT demonstrated superior performance to this index, the inclusion of anthropometric and hormonal markers could enhance the predictive value for identifying adolescents at risk of MASLD in high-risk groups. Future studies should consider incorporating these factors as they are easily applicable in the clinical setting and may provide a more comprehensive assessment tool for early intervention.

Strengths and limitations

The strengths of this study include its diverse sample of children recruited from multiple geographic regions of the United States. In addition, the use of 1 of 2 accurate diagnostic methods—MRI-PDFF or ¹H-MRS—increased our confidence in the subject’s classification of MASLD. Furthermore, as mentioned above, advanced LC-MS/MS techniques allowed us to accurately measure metabolically active insulin molecules for detecting MASLD in youth.

Our investigation was limited by its small sample size and the imbalance between the number of participants in the MASLD and non-MASLD groups, males versus females, and Hispanic race/ethnicity versus other races. These factors limit the generalizability of our findings, given that MASLD is more prevalent in males of Hispanic descent.⁵⁰ Moreover, our patient population was primarily drawn from studies focusing on children at high risk for or with a known diagnosis of MASLD, which could also restrict the generalizability of our conclusions.

Of note, our study included data from youth with increased risk for IR due to obesity, adolescence, PCOS, and Hispanic ethnicity. IR appears to be critical for the metabolic component of MASLD. However, further studies that evaluate the predictive nature of fasting intact insulin in patients with less IR may offer added insight for identifying MASLD in at-risk youth.

In this study, we did not have complete data for the standard radioimmunoassay-measured fasting insulin alongside intact insulin for all participants, preventing us from including this analysis. In a prior study, Bril et al²⁴ found that radioimmunoassay was inferior to intact insulin measured by LC-MS/MS for identifying MASLD in adults without T2DM (p = 0.007), suggesting that intact insulin measurements may enhance the diagnostic performance of fasting insulin in detecting MASLD. Furthermore, the LC-MS/MS assay used in our study for insulin and C-peptide is traceable to the peptide content and provides standardized measurements for both molecules.^23,26

Lastly, our MASLD group had significantly more children with obesity, although the median BMI percentile among both the MASLD and non-MASLD groups was >95%. Larger studies involving more diverse populations are necessary to confirm our findings. In addition, independent validation of intact insulin, both alone and in combination with ALT, for identifying MASLD in an external pediatric cohort is warranted.

CONCLUSIONS

As more therapeutic options emerge, early recognition of children at risk for MASLD is crucial. Prompt detection helps to identify patients who may benefit from further workup and enables adequate treatment to delay disease progression. Our study indicates that similar to adults, measures of fasting intact insulin alone and in combination with ALT (intact insulin × ALT) are strongly associated with MASLD in children and adolescents. Although further validation in larger and more diverse populations is needed, our findings offer promising support for the use of fasting intact insulin and ALT levels as a simple, noninvasive method for diagnosing MASLD in pediatric patients.

Supplementary Material

hc9-8-e0582-s001.doc ^{(94.5KB, doc)}