Abstract
Background
The relationship between pancreatic fat on imaging and metabolic co-morbidities has not been established in pediatrics. We sought to investigate the relationship between pancreatic fat measured by MRI and endocrine/exocrine dysfunctions along with the metabolic co-morbidities in a cohort of children.
Objective
To investigate relationships between pancreatic fat quantified by MRI and endocrine and exocrine conditions and metabolic co-morbidities in a cohort of children.
Materials and Methods
This was a retrospective review of pediatric patients (n = 187) who had a clinically indicated MRI examination between May 2018 and February 2020. After 51 patients without useable imaging data were excluded, the remaining 136 subjects comprised the study sample. Laboratory studies were assessed if collected within 6 months of MRI and patient charts were reviewed for demographic and clinical information. MRI proton density fat fraction (PDFF) sequence had been acquired according to manufacturer’s specified parameters at a slice thickness of 3 mm. Two blinded radiologists independently collected PDFF data.
Results
The median age at MRI was 12.1 (IQR: 9.0–14.8) years and the majority of patients were Caucasian (79%), followed by African American and Hispanic at 12% and 11% respectively. There was a higher median pancreas fat fraction in patients with exocrine conditions (chronic pancreatitis or exocrine insufficiency) compared to those without (3.5% vs 2.2%, p = 0.03). There was also a higher median fat fraction in the head of pancreas in patients with endocrine insufficient conditions (insulin resistance, pre-diabetes, type 1 and type 2 diabetes) compared to those without endocrine insufficiency when excluding patients with active acute pancreatitis (3.5% vs 2.0%, p = 0.04). Patients with BMI > 85% had higher mean fat fraction compared to patients with BMI ≤ 85% (head: 3.8 vs 2.4%, p = 0.01; body: 3.8 vs 2.5%, p = 0.005; tail: 3.7 vs 2.7%, p = 0.049; overall pancreas fat fraction: 3.8 vs 2.6%, p = 0.002).
Conclusion
Pancreas fat is elevated in patients with BMI > 85% and in those with exocrine and endocrine insufficiencies.
Keywords: Pancreas fat, MRI proton density fat fraction, Exocrine insufficiency, Endocrine insufficiency
Graphical Abstract
Introduction
Pediatric obesity is an increasingly prevalent health concern and leads to metabolic consequences [1]. The relationship between obesity, visceral fat deposition, and metabolic consequences is well described [2–4]. Although obesity plays a key role in the development of insulin resistance and type 2 diabetes (T2DM), the specific mechanisms underlying this progression are still largely unknown [5]. The relationship between liver fat deposition and progression to disease has been well described [6]. However, the relationship between obesity, pancreatic fat, and clinical disease is less well understood, particularly in pediatrics. A meta-analysis of studies evaluating pancreatic imaging findings showed increased fat deposition and lower pancreatic volume in adults with T2DM when compared to age-matched controls [7]. However, this has not been shown to be consistent in pediatric studies as two recent studies evaluating pediatric patients with nonalcoholic fatty liver disease (NAFLD) did not find a difference in pancreas fat content and occurrence of pre-diabetes or T2DM [8, 9].
There are various imaging modalities to evaluate pancreas fat including ultrasound, CT, and MRI. Ultrasound is less accurate as the fat assessment is qualitative [10]. CT is less preferred in children due to ionizing radiation exposure. Magnetic resonance imaging proton density fat fraction (MR-PDFF) is a technique that accurately quantifies tissue fat and correlates well with histology-proven fat content in various body sites [11]. MRI has been shown to be an effective modality for quantifying pancreatic fat fraction [12].
As the significance of pancreatic fat in the pediatric population has not been well established, we sought to investigate relationships between pancreatic fat quantified by MRI and endocrine and exocrine conditions and metabolic co-morbidities in a cohort of children.
Methods
This was an IRB approved (2020–0154) retrospective review of pediatric patients age < 18 years (n = 187) who had undergone clinically indicated MRI examinations of the abdomen that included an MRI proton density fat fraction (PDFF) sequence performed at Cincinnati Children’s Hospital Medical Center between May 2018 and February 2020. The total subjects with images that had available fat fraction data were 136, as 51 patients with no useable fat fraction data were excluded. Patients were identified by querying the electronic medical record using Illuminate (Illuminate; Overland Park, KS). Metabolic studies and clinical laboratory values (including HgbA1c, mixed meal studies, lipids, LFTs, lipase, fecal elastase) were collected if done within 6 months of MRI. Patient charts were also reviewed for diagnosis codes relating to pancreatitis, diabetes, nonalcoholic steatohepatitis (NASH), and cystic fibrosis. Demographic and clinical parameters were collected including age, sex, ethnicity, race, blood pressure, and BMI. Overweight was defined as BMI > 85% [13]. Endocrine insufficiency included patients with insulin resistance, pre-diabetes, and type 1 and type 2 diabetes. Diabetes was defined as per the ADA criteria [14, 15]. Pre-diabetes was defined as HgbA1c between 5.7 and 6.4%, fasting glucose > 100 mg/dL, and 2-h post-oral glucose tolerance test > 140 mg/dL, and insulin resistance was defined using diagnosis codes at clinical visits based upon presence of acanthosis nigricans or laboratory values consistent with insulin resistance. Pancreatitis was defined as per INSPPIRE criteria as follows: acute pancreatitis diagnosis requires 2 of 3 criteria, abdominal pain suggestive of acute pancreatitis (acute onset particularly in epigastric region); amylase and/or lipase at least 3 times greater than the upper limit of normal and imaging findings characteristic of acute pancreatitis. Acute recurrent pancreatitis diagnosis requires at least 2 distinct episodes of acute pancreatitis along with either complete resolution of pain (≥ 1 month pain free between episodes) or normalization of serum amylase/lipase before subsequent acute pancreatitis is diagnosed (also with resolution of pain irrespective of time interval). Chronic pancreatitis diagnosis requires 1 of 3 criteria: abdominal pain consistent with pancreatic origin and imaging findings suggestive of pancreatic damage; exocrine pancreatic insufficiency and suggestive imaging findings; endocrine pancreatic insufficiency and suggestive imaging findings; or, a pancreas biopsy specimen demonstrating histopathologic features of chronic pancreatitis [15].
Pancreas, liver, and subcutaneous and visceral fat quantification
MRI of the abdomen had been performed according to various clinical protocols. A variety of MRI scanners were used including Philips Ingenia 1.5 T and 3 T and GE 750W 3 T. The proton density fat fraction (PDFF) sequence was acquired according to manufacturer’s specified parameters at a slice thickness of 3 mm.
Two board-certified pediatric radiologists each independently reviewed the MRI examinations and placed a single ovoid region of interest (ROIs) in each of the head, body, and tail of the pancreas on the fat fraction parametric map to quantify MRI PDFF. Measurements were made in our clinical PACS (Merge PACS, IBM Watson Health). Observers were blinded to each other’s measurements and ROIs were adjusted to be as large as possible while remaining entirely within the pancreas. Values were rounded to the nearest whole number and overall fat fraction was calculated as a simple mean of segmental measurements. An example of patient MRI PDFF image is shown in Fig. 1.
Fig. 1.
Axial proton density fat fraction map a and water b images derived from the same 6-point Dixon acquisition in an 11-year-old boy with pre-diabetes and impaired glucose tolerance. Images represent sampling of pancreas fat fraction in the body and tail (ovoid regions of interest). PDFF measurements are derived from the PDFF map. Regions of interest are propagated to the water image to show sampling locations. In this case, overall pancreas fat fraction was 5% as measured by both observers (3%/6%, 6%/3%, and 6%/6% in the head, body, and tail of the pancreas respectively for R1/R2)
Liver fat fraction measurements were made by a single observer by placing equal size large ovoid ROIs in an area of the right hepatic lobe free from vessels on three separate images per patient. Liver PDFF was then calculated as the simple mean on these three measurements.
Fat thickness measurements were made by a single observer as linear measurements at the L3 vertebral level between the skin surface and external surface of the rectus muscle (subcutaneous fat thickness) and between the internal surface of the rectus muscle and the anterior margin of the L3 vertebral body (visceral fat thickness).
Statistical analysis
Data were analyzed using SAS®, version 9.4 (SAS Institute, Cary, NC). Depending on distributions, continuous data were summarized as either means ± standard deviations (SD) or medians with interquartile ranges (IQR: 25th–75th percentiles), while categorical data were summarized as frequency counts and percentages. For categorical data, chi-square or Fisher’s exact tests were used, as appropriate, for group comparisons. T-tests, Mann–Whitney-Wilcoxon, or Kruskal–Wallis tests were used, as appropriate, for continuous data comparisons. Pearson correlations were also run between the two radiologist readers. Since there were 2 readers, the mean of the reader values was used for each site (head, body, and tail) for each patient. The overall pancreas fat fraction was calculated as the mean of the 3 pancreas sites (head, body, and tail) for each patient. A p-value of < 0.05 was considered statistically significant.
Results
Overall cohort clinical characteristics
There were 136 patients who underwent MRI with available fat fraction data included in this study; 8 of those did not have pancreas fat data available but had liver and subcutaneous and visceral fat data. Demographic and clinical information is shown in Table 1. The median age at MRI was 12.1 (IQR: 9.0–14.8) years and the majority of patients were Caucasian (79%), followed by African American and Hispanic at 12% and 11% respectively. There were 30 patients (22%) with glycemic abnormalities including pre-diabetes, impaired fasting glucose, insulin resistance, and type 2 and type 1 diabetes. Within our cohort, 18% had acute pancreatitis without recurrence, 22% had acute recurrent pancreatitis, and 21% had chronic pancreatitis. Thirty-eight percent of our cohort had BMI > 85%. There were two patients diagnosed with cystic fibrosis and one patient with Shwachman-Diamond syndrome.
Table 1.
Patient demographics and clinical characteristics. Data presented as median (25th–75th percentile) or n (%)
| Useable procedures (N = 136) | |
|---|---|
|
| |
| Procedure age (years) | 12.1 (9.0–14.8) n = 136 |
| Sex (female) | 76/136 (56%) |
| Race | |
| White/Caucasian | 102/129 (79%) |
| African American | 16/129 (12%) |
| Other | 11/129 (9%) |
| Ethnicity (non-Hispanic) | 119/133 (89%) |
| Height percentile | 53.7 (18.0–75.5) n = 120 |
| Weight percentile | 68.9 (30.5–93.9) n = 133 |
| BMI (kg/m2) | 20.7 (17.4–24.2) n = 116 |
| BMI percentile | 75.0 (42.9–95.9) n = 116 |
| BMI > 85th percentile | 44/116 (38%) |
| Pre-diabetes/impaired glucose tolerance | 19/136 (14%) |
| Insulin resistance | 2/136 (1.5%) |
| Type 2 diabetes | 1/136 (0.7%) |
| Type 1 diabetes | 4/136 (2.9%) |
| Other diabetes | 4/136 (2.9%) |
| Hypertension | 10/126 (8%) |
| Pancreatitis diagnosis | |
| Acute pancreatitis | 24/136 (18%) |
| Acute recurrent pancreatitis | 30/136 (22%) |
| Chronic pancreatitis | 29/136 (21%) |
| None | 53/136 (39%) |
| Cystic fibrosis | 2/136 (1.5%) |
| Shwachman-Diamond | 1/136 (0.7%) |
| NASH | 3/136 (2.2%) |
| Exocrine insufficiency (mcg/g) | |
| Normal (> 200) | 20/35 (57%) |
| Moderate (100–200) | 9/35 (26%) |
| Severe (< 100) | 6/35 (17%) |
| Exocrine vs endocrine insufficiency | |
| Exocrine only | 63/136 (46%) |
| Endocrine only | 10/136 (7%) |
| Both | 20/136 (15%) |
| None | 43/136 (32%) |
Agreement of pancreas fat quantification between readers
Two radiologists independently performed the fat fraction measurements. Figure 2 depicts the correlation of various pancreas fat sites (head, body, tail) between reader 1 (R1) and reader 2 (R2). The head showed the highest degree of correlation between the readers (r = 0.74), followed by the body (r = 0.52), and the tail showed the lowest correlation (r = 0.44) between the two radiologists reading the images.
Fig. 2.
Pearson correlations between reader 1 (R1) and reader 2 (R2) at the various pancreas sites measured including overall pancreas fat fraction a, fat fraction of the pancreas head b, fat fraction of the pancreas body c, and fat fraction of the pancreas tail d
Fat quantification on imaging and clinical outcomes
There was no statistical difference found between pancreas (p = 0.66), liver (p = 0.25), visceral fat (p = 0.26), or subcutaneous fat (p = 0.56) deposition when comparing patients with both endocrine/exocrine, endocrine only, exocrine only, versus no conditions (Table 2). However, when evaluating only those patients with exocrine pancreatic insufficiency (EPI) or chronic pancreatitis (CP) (n = 6 had chronic pancreatitis and EPI, n = 8 had chronic pancreatitis without EPI, and n = 15 had chronic pancreatitis and unknown EPI status) compared to those who had no exocrine or endocrine problems (Table 3), there was a statistically significant difference in median overall pancreas fat fraction (3.5% in CP/EPI vs 2.2% in the group without CP/EPI, p = 0.03). Differences in pancreas fat fraction in the head (3.0% in CP/EPI vs 2.0% in the group without CP/EPI, p = 0.06), body (2.5% in CP/EPI vs 2.5% in the group without CP/EPI, p = 0.65), and tail (3.0% in CP/EPI vs 2.0% in the group without CP/EPI, p = 0.17) did not reach statistical significance. When evaluating patients who had endocrine insufficient conditions and excluding those with active acute pancreatitis (Table 4), there was a statistically significant difference in median pancreas fat fraction in the head (3.5% vs 2.0%, p = 0.04). There was no statistical difference found between fat deposition at any site in relation to race or ethnicity in our cohort (Supplemental Tables 1 and 2).
Table 2.
Fat deposition on imaging and endocrine and exocrine insufficiency. Data presented as median (25th–75th percentile)
| Both endocrine/exocrine* (N = 20) | Endocrine only* (N = 10) | Exocrine only* (N = 63) | None* (N = 43) | p-value | |
|---|---|---|---|---|---|
|
| |||||
| Procedure age (years) | 12.0 (9.6–14.6), n = 20 | 9.9 (8.9–15.4), n = 10 | 12.0 (7.8–14.5), n = 63 | 13.5 (10.2–16.5), n = 43 | 0.09 |
| Sex (female) | 9/20 (45%) | 1/10 (10%) | 43/63 (68%) | 23/43 (53%) | 0.003 |
| Height percentile | 33.6 (11.2–78.7), n = 19 | 45.6 (20.3–89.8), n = 8 | 54.1 (18.4–76.0), n = 56 | 57.9 (31.8–70.1), n = 37 | 0.83 |
| Weight percentile | 73.7 (32.1–96.9), n = 20 | 95.4 (36.3–99.4), n = 9 | 69.7 (30.5–90.2), n = 62 | 62.6 (29.7–86.5), n = 42 | 0.56 |
| BMI percentile | 75.0 (55.3–96.8), n = 19 | 96.5 (45.1–99.5), n = 8 | 77.2 (50.6–94.2), n = 53 | 64.8 (24.6–93.2), n = 36 | 0.37 |
| Pancreas fat fraction (head, %) | 3.0 (1.5–3.5), n = 17 | 3.5 (3.0–4.5), n = 9 | 2.5 (1.0–5.0), n = 59 | 2.0 (1.0–3.0), n = 41 | 0.31 |
| Pancreas fat fraction (body, %) | 3.0 (2.0–5.0), n = 15 | 2.0 (1.5–4.0), n = 9 | 2.5 (1.5–4.0), n = 56 | 2.5 (1.5–3.5), n = 41 | 0.83 |
| Pancreas fat fraction (tail, %) | 2.0 (1.0–4.0), n = 15 | 3.0 (1.5–4.5), n = 9 | 3.0 (1.0–4.5), n = 59 | 2.0 (1.0–4.0), n = 41 | 0.78 |
| Overall pancreas fat fraction (%) | 2.7 (2.0–4.0), n = 17 | 3.2 (2.2–3.8), n = 9 | 2.5 (1.8–4.2), n = 61 | 2.2 (1.5–3.3), n = 41 | 0.66 |
| Subcutaneous fat thickness (cm) | 15.3 (10.5–25.5), n = 20 | 23.2 (8.5–39.5), n = 10 | 16.8 (7.3–28.0), n = 63 | 13.8 (6.5–24.4), n = 42 | 0.56 |
| Visceral fat thickness (mm) | 57.7 (47.7–78.3), n = 20 | 56.4 (35.6–79.0), n = 10 | 48.8 (42.1–59.3), n = 63 | 53.4 (45.9–67.3), n = 42 | 0.26 |
| Liver fat fraction (%) | 3.2 (2.2–4.5), n = 20 | 4.2 (2.0–11.7), n = 10 | 3.0 (2.0–4.3), n = 63 | 3.7 (3.0–5.0), n = 42 | 0.25 |
Endocrine insufficiency—pre-diabetes/impaired glucose tolerance, insulin resistance, type 1 or 2 diabetes, or other diabetes
Exocrine insufficiency—exocrine pancreatic insufficiency (< 100), acute pancreatitis, acute recurrent pancreatitis, or chronic pancreatitis
None—no endocrine or exocrine insufficiency
p-values which have reached statistical significance (< 0.05) are bolded
Table 3.
Fat deposition on imaging and chronic pancreatitis/exocrine insufficiency vs none. Data presented as median (25th–75th percentile)
| Chronic pancreatitis/exocrine insufficiency (N = 29) | None* (N = 43) | p-value | |
|---|---|---|---|
|
| |||
| Procedure age (years) | 10.9 (8.1–13.8), n = 29 | 13.5 (10.2–16.5), n = 43 | 0.02 |
| Sex (female) | 17/29 (59%) | 23/43 (53%) | 0.67 |
| Height percentile | 50.5 (16.7–76.3), n = 27 | 57.9 (31.8–70.1), n = 37 | 0.72 |
| Weight percentile | 71.3 (32.7–96.9), n = 29 | 62.6 (29.7–86.5), n = 42 | 0.33 |
| BMI percentile | 79.7 (67.1–96.8), n = 27 | 64.8 (24.6–93.2), n = 36 | 0.08 |
| Pancreas fat fraction (head, %) | 3.0 (2.0–5.0), n = 25 | 2.0 (1.0–3.0), n = 41 | 0.06 |
| Pancreas fat fraction (body, %) | 2.5 (1.5–5.0), n = 23 | 2.5 (1.5–3.5), n = 41 | 0.65 |
| Pancreas fat fraction (tail, %) | 3.0 (1.5–4.5), n = 23 | 2.0 (1.0–4.0), n = 41 | 0.17 |
| Overall pancreas fat fraction (%) | 3.5 (2.3–4.3), n = 25 | 2.2 (1.5–3.3), n = 41 | 0.03 |
| Subcutaneous fat thickness (cm) | 17.5 (11.7–27.2), n = 29 | 13.8 (6.5–24.4), n = 42 | 0.29 |
| Visceral fat thickness (mm) | 51.9 (46.7–63.5), n = 29 | 53.4 (45.9–67.3), n = 42 | 0.95 |
| Liver fat fraction (%) | 3.3 (2.0–4.0), n = 29 | 3.7 (3.0–5.0), n = 42 | 0.30 |
None—no endocrine or exocrine insufficiency
p-values which have reached statistical significance (< 0.05) are bolded
Table 4.
Fat deposition on imaging and endocrine insufficiency excluding those with active acute pancreatitis. Data presented as median (25th–75th percentile)
| Any Endocrine insufficiency* (no active acute pancreatitis), N = 17 | None*, (no active acute pancreatitis), N = 41 | p-value | |
|---|---|---|---|
|
| |||
| Procedure age (years) | 10.1 (9.1–15.0), n = 17 | 13.6 (10.2–16.5), n = 41 | 0.30 |
| Sex (female) | 4/17 (24%) | 21/41 (51%) | 0.08 |
| Height percentile | 58.1 (16.0–85.6), n = 15 | 58.3 (31.8–70.7), n = 35 | 0.66 |
| Weight percentile | 83.0 (49.6–99.4), n = 17 | 62.6 (30.0–87.9), n = 40 | 0.06 |
| BMI percentile | 80.6 (48.6–99.4), n = 15 | 64.8 (24.3–95.0), n = 34 | 0.08 |
| Pancreas fat fraction (head, %) | 3.5 (3.0–3.5), n = 17 | 2.0 (1.0–3.0), n = 41 | 0.04 |
| Pancreas fat fraction (body, %) | 3.0 (2.0–5.0), n = 17 | 2.5 (1.5–3.5), n = 41 | 0.33 |
| Pancreas fat fraction (tail, %) | 1.5 (1.0–4.5), n = 17 | 2.0 (1.0–4.0), n = 41 | 0.96 |
| Overall pancreas fat fraction (%) | 3.2 (2.5–4.0), n = 17 | 2.2 (1.5–3.3), n = 41 | 0.08 |
| Subcutaneous fat thickness (cm) | 18.8 (11.7–39.5), n = 17 | 13.8 (7.0–25.9), n = 40 | 0.13 |
| Visceral fat thickness (mm) | 50.0 (42.1–69.7), n = 17 | 53.4 (45.5–67.4), n = 40 | 0.85 |
| Liver fat fraction (%) | 4.3 (3.0–14.3), n = 17 | 3.7 (3.0–5.0), n = 40 | 0.15 |
Endocrine insufficiency—pre-diabetes/impaired glucose tolerance, insulin resistance, type 1 or 2 diabetes, or other diabetes
None—no endocrine or exocrine insufficiency
p-values which have reached statistical significance (< 0.05) are bolded
Fat quantification on imaging and overweight body habitus
Table 5 depicts the relationship between fat deposition on imaging and BMI status. Overweight subjects with BMI > 85% for age had higher mean fat deposition in all sites within the pancreas compared to those with BMI ≤ 85% (head: 3.8 vs 2.4%, p = 0.01; body: 3.8 vs 2.5%, p = 0.005; tail: 3.7 vs 2.7%, p = 0.049; overall pancreas fat fraction: 3.8 vs 2.6%, p = 0.002). Patients with BMI > 85% also had higher median subcutaneous fat thickness (29.7 vs 11.7 cm, p < 0.0001), visceral fat thickness (60.3 vs 48.6 mm, p = 0.001), and liver fat fraction (4.7 vs 3.0%, p = 0.001).
Table 5.
Fat deposition on imaging and BMI > 85%. Data presented as median (25th–75th percentile) or mean ± SD or n (%)
| BMI > 85%, N = 44 | BMI ≤ 85%, N = 72 | p-value | |
|---|---|---|---|
|
| |||
| Procedure age (years) | 12.1 (9.7–14.6), n = 44 | 12.8 (9.5–15.3), n = 72 | 0.81 |
| Sex (female) | 22/44 (50%) | 40/72 (56%) | 0.56 |
| Height percentile | 63.0 (31.8–84.9), n = 44 | 46.8 (16.2–67.7), n = 72 | - |
| Weight percentile | 97.5 (89.7–99.0), n = 44 | 44.7 (14.0–71.1), n = 72 | - |
| BMI percentile | 97.0 (94.4–99.0), n = 44 | 52.8 (23.7–73.0), n = 72 | - |
| Pancreas fat fraction (head, %) | 3.8 ± 2.9, n = 41 | 2.4 ± 1.9, n = 66 | 0.01 |
| Pancreas fat fraction (body, %) | 3.8 ± 2.4, n = 38 | 2.5 ± 2.0, n = 65 | 0.005 |
| Pancreas fat fraction (tail, %) | 3.7 ± 2.7, n = 40 | 2.7 ± 2.5, n = 65 | 0.049 |
| Overall pancreas fat fraction (%) | 3.8 ± 2.2, n = 42 | 2.6 ± 1.7, n = 67 | 0.002 |
| Subcutaneous fat thickness (cm) | 29.7 (20.9–38.3), n = 44 | 11.7 (5.7–16.9), n = 71 | < 0.0001 |
| Visceral fat thickness (mm) | 60.3 (48.3–70.9), n = 44 | 48.6 (42.1–54.7), n = 71 | 0.001 |
| Liver fat fraction (%) | 4.7 (2.3–9.2), n = 44 | 3.0 (2.3–4.0), n = 71 | 0.001 |
p-values which have reached statistical significance (< 0.05) are bolded
Increased pancreas fat deposition and clinical outcomes
Table 6 shows clinical factors in relation to pancreas fat deposition any site > 6%. Patients with increased pancreas fat deposition (head, body, or tail > 6%) had higher BMI percentile compared to those with less pancreatic fat deposition (96.1% vs 70.0%, p = 0.001). There was a statistical difference in subcutaneous fat thickness between those with increased pancreas fat > 6% and those with ≤ 6% pancreas fat (26.7 cm vs 15.8 cm, p = 0.01), but no difference found in visceral fat (54.2 mm vs 49.6 mm, p = 0.09) and liver fat fraction (4.2% vs 3.3%, p = 0.08). There was no difference found in age, sex, race, ethnicity, exocrine condition, endocrine condition, fasting glucose, HgbA1c, HDL, LDL, triglycerides, or liver function studies between those who had increased (> 6%) pancreas fat and those with lower (≤ 6%) pancreas fat.
Table 6.
Association of elevated pancreas fat deposition and clinical co-morbidities. Data presented as median (25th–75th percentile) or mean ± SD or n (%)
| Pancreas fat > 6%*, N = 24 | Pancreas fat ≤ 6%, N = 104 | p-value | |
|---|---|---|---|
|
| |||
| Procedure age (years) | 11.5 (7.8–16.1), n = 24 | 12.2 (9.2–14.8), n = 104 | 0.84 |
| Sex (female) | 13/24 (54%) | 58/104 (56%) | 0.89 |
| Race | 0.26 | ||
| White/Caucasian | 16/22 (73%) | 80/99 (81%) | |
| African American | 5/22 (23%) | 10/99 (10%) | |
| Other | 1/22 (5%) | 9/99 (9%) | |
| Ethnicity (non-Hispanic) | 22/24 (92%) | 90/104 (89%) | 1.00 |
| Height percentile | 50.4 (16.0–67.0), n = 21 | 58.2 (18.6–79.5), n = 92 | 0.45 |
| Weight percentile | 77.8 (51.4–98.9), n = 24 | 67.1 (28.3–89.3), n = 102 | 0.048 |
| BMI percentile | 96.1 (76.6–98.9), n = 20 | 70.0 (33.8–91.4), n = 89 | 0.001 |
| Pancreatitis diagnosis | 0.19 | ||
| Acute pancreatitis | 2/24 (8%) | 21/104 (20%) | |
| Acute recurrent pancreatitis | 4/24 (17%) | 26/104 (25%) | |
| Chronic pancreatitis | 8/24 (33%) | 17/104 (16%) | |
| None | 10/24 (42%) | 40/104 (38%) | |
| Pancreas fat fraction (head, %) | 5.2 ± 3.5, n = 24 | 2.3 ± 1.6, n = 102 | - |
| Pancreas fat fraction (body, %) | 5.8 ± 2.9, n = 23 | 2.4 ± 1.4, n = 98 | - |
| Pancreas fat fraction (tail, %) | 6.5 ± 2.8, n = 24 | 2.2 ± 1.6, n = 100 | - |
| Overall pancreas fat fraction (%) | 5.9 ± 1.9, n = 24 | 2.3 ± 1.2, n = 104 | - |
| Subcutaneous fat thickness (cm) | 26.7 (11.7–38.6), n = 24 | 15.8 (6.7–26.7), n = 103 | 0.01 |
| Visceral fat thickness (mm) | 54.2 (47.2–69.6), n = 24 | 49.6 (42.1–65.3), n = 103 | 0.09 |
| Liver fat fraction (%) | 4.2 (3.0–5.7), n = 24 | 3.3 (2.0–4.3), n = 103 | 0.08 |
| Exocrine vs endocrine insufficiency | 0.94 | ||
| Exocrine only | 11/24 (46%) | 50/104 (48%) | |
| Endocrine only | 1/24 (4%) | 8/104 (8%) | |
| Both | 3/24 (13%) | 14/104 (13%) | |
| None | 9/24 (38%) | 32/104 (31%) | |
| Family history of diabetes mellitus | 3/8 (38%) | 14/27 (52%) | 0.69 |
| Suspected pancreatitis at home | 3/8 (38%) | 11/27 (41%) | 1.00 |
| HgbA1c | 4.9 (4.8–5.3), n = 10 | 5.1 (4.9–5.3), n = 38 | 0.33 |
| Fasting glucose | 94.5 (85.0–102.0), n = 18 | 89.0 (81.0–102.0), n = 91 | 0.64 |
| Fasting C-peptide | 0.8 (0.3–1.1), n = 3 | 1.1 (0.7–1.9), n = 22 | 0.34 |
| Total cholesterol | 154.0 (125.0–191.0), n = 9 | 146.0 (129.0–169.0), n = 37 | 0.47 |
| hdl | 36.5 (32.0–41.5), n = 8 | 41.5 (36.0–50.0), n = 34 | 0.20 |
| < 40 | 5/8 (63%) | 14/34 (41%) | |
| ≥ 40 | 3/8 (38%) | 20/34 (59%) | |
| LDL | 87.0 (75.5–100.5), n = 8 | 82.0 (63.0–99.0), n = 34 | 0.59 |
| > 130 | 0/8 (0%) | 3/34 (9%) | |
| Triglycerides | 96.0 (63.0–214.0), n = 11 | 90.0 (66.0–147.0), n = 54 | 0.78 |
| > 150 | 3/11 (27%) | 13/54 (24%) | |
| AST | 30.0 (26.0–47.0), n = 21 | 29.0 (21.0–49.0), n = 91 | 0.44 |
| ALT | 30.0 (16.0–91.0), n = 21 | 26.0 (16.0–64.0), n = 91 | 0.54 |
Pancreas fat > 6%—any site (head, body, or tail fat fraction > 6%)
p-values which have reached statistical significance (< 0.05) are bolded
Discussion
To our knowledge, this is the first pediatric study examining the relationship between pancreas fat deposition and endocrine/exocrine co-morbidities as well as obesity and other metabolic measures. We have shown that pediatric patients with chronic pancreatitis and exocrine insufficiency have higher pancreatic fat in comparison to patients with no exocrine conditions. We also found that patients with endocrine conditions had higher fat deposition within the pancreas head, when excluding those with active acute pancreatitis. We chose to exclude the patients with active acute pancreatitis in the comparison of endocrine conditions as the acute inflammatory reaction may mask fat reading results [16]. Patients with overweight body habitus had higher fat deposition in all sites measured, including pancreas, liver, and subcutaneous and visceral fat. Another key finding was the challenge of measuring pancreatic fat and the variability that can be seen between readers. In our study, the pancreas head was the site of best correlation and the tail had the weakest correlation between reader 1 and reader 2.
As expected, patients with overweight body habitus had higher deposition of pancreatic fat compared to those with lower BMI, which is in line with prior pediatric studies [9, 17–19]. A recent study in adults demonstrated a decrease in pancreas fat deposition after weight loss surgery, suggesting that pancreas fat deposition is closely tied to anthropometric measures [20]. Patients with diagnoses of endocrine conditions, including insulin resistance, pre-diabetes, type 2 diabetes, or type 1 diabetes, were found to have significantly higher pancreas fat deposition in the head only when compared to those without any endocrine dysfunction. Prior studies in adults with type 2 diabetes have also shown a preferential loss of islets in the head of the pancreas which may explain this finding [21, 22]. A recent study in overweight adolescents described an association between fatty pancreas on MRI and insulin resistance/beta cell dysfunction [23]. However, another study in adolescents found no relationship between pancreatic fat deposition and beta cell dysfunction, except in those who were carriers of G319S mutation of the HNF-1 alpha gene [24]. Another recent study in Latino adolescents found that pancreatic fat was related to fasting glycemia but not with HgbA1c values [25]. The relationship between pancreas fat and youth onset diabetes is still an area of inconclusive findings and warrants further prospective evaluation. Our study found no difference between fat deposition and race or ethnicity. However, a prior study in adolescents demonstrated that pancreas fat deposition in African American youth was predicative of the development of pre-diabetes [26]. Exocrine disease was also found to be associated with higher pancreas fat deposition in our study. A prospective study in a cohort of healthy adults found that a higher presence of pancreas fat as visualized on ultrasound was associated with a higher odd of developing subclinical chronic pancreatitis [27]. There is a paucity of pediatric data evaluating exocrine disease and fatty pancreas, which we believe warrants further study.
Our study while adding novel findings to the field has limitations that include the retrospective study design which may have introduced bias, the 6-month interval between MRI and clinical studies collected, the variation in MRI scanners used, some subjects with missing laboratory data, and the small sample size of patients with endocrine insufficient conditions. Due to our small numbers, we had to combine all patients with any endocrine abnormality in our analysis, which may have masked some of the results relevant to different types of diabetes. It could also be considered a limitation that the fat fraction measurements from the two radiologists were in many cases incongruent, which may have been related to placement of the ROI in the context of distorted anatomy related to pancreatic atrophy. Main strengths are that the chronic pancreatitis and exocrine insufficiency patients had a relatively larger sample size, the analysis of various clinical co-morbidities, and having two independent radiologists as readers allowed us to identify sites within the pancreas demonstrating the highest correlation between readers.
In conclusion, pancreas fat was found to be elevated in patients with increased body habitus and in patients who have exocrine pancreas disease (chronic pancreatitis and exocrine insufficiency). Elevated pancreas fat fraction in the head was found to be associated with endocrine disease. Future studies should focus on prospectively evaluating the progression of pancreas fat within the various sites (head, body, and tail) in pediatric subjects with endocrine and exocrine conditions. Understanding the implications of pancreatic fat deposition will help guide clinicians in identifying patients at risk for developing future endocrine and exocrine disease outcomes.
Supplementary Material
Footnotes
Conflicts of interest The authors declare no competing interest.
Ethics approval This study was approved (2020–0154) by the institution’s IRB.
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s00247-023-05669-8.
Data Availability
The data that support the findings of this study are available from the corresponding author, S.E.S., upon reasonable request.
References
- 1.Styne DM, Arslanian SA, Connor EL et al. (2017) Pediatric obesity-assessment, treatment, and prevention: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 102:709–757 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Staiano AE, Katzmarzyk PT (2012) Ethnic and sex differences in body fat and visceral and subcutaneous adiposity in children and adolescents. Int J Obes (Lond) 36:1261–1269 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jung JH, Jung MK, Kim KE et al. (2016) Ultrasound measurement of pediatric visceral fat thickness: correlations with metabolic and liver profiles. Ann Pediatr Endocrinol Metab 21:75–80 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lopes HF, Corrêa-Giannella ML, Consolim-Colombo FM, Egan BM (2016) Visceral adiposity syndrome. Diabetol Metab Syndr 8:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Valaiyapathi B, Gower B, Ashraf AP (2020) Pathophysiology of type 2 diabetes in children and adolescents. Curr Diabetes Rev 16:220–229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Pierantonelli I, Svegliati-Baroni G (2019) Nonalcoholic fatty liver disease: basic pathogenetic mechanisms in the progression from NAFLD to NASH. Transplantation 103:e1–e13 [DOI] [PubMed] [Google Scholar]
- 7.Garcia TS, Rech TH, Leitão CB (2017) Pancreatic size and fat content in diabetes: a systematic review and meta-analysis of imaging studies. PLoS ONE 12:e0180911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Trout AT, Hunte DE, Mouzaki M et al. (2019) Relationship between abdominal fat stores and liver fat, pancreatic fat, and metabolic comorbidities in a pediatric population with non-alcoholic fatty liver disease. Abdom Radiol (NY) 44:3107–3114 [DOI] [PubMed] [Google Scholar]
- 9.Kim J, Albakheet SS, Han K et al. (2021) Quantitative MRI assessment of pancreatic steatosis using proton density fat fraction in pediatric obesity. Korean J Radiol 22:1886–1893 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sepe PS, Ohri A, Sanaka S et al. (2011) A prospective evaluation of fatty pancreas by using EUS. Gastrointest Endosc 73:987–993 [DOI] [PubMed] [Google Scholar]
- 11.Idilman IS, Tuzun A, Savas B et al. (2015) Quantification of liver, pancreas, kidney, and vertebral body MRI-PDFF in non-alcoholic fatty liver disease. Abdom Imaging 40:1512–1519 [DOI] [PubMed] [Google Scholar]
- 12.Majumder S, Philip NA, Takahashi N et al. (2017) Fatty pancreas: should we be concerned? Pancreas 46:1251–1258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Barlow SE, Committee E (2007) Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics 120(Suppl 4):S164–192 [DOI] [PubMed] [Google Scholar]
- 14.Association AD (2021) 2. Classification and diagnosis of diabetes. Diabetes Care 44:S15–S33 [DOI] [PubMed] [Google Scholar]
- 15.Morinville VD, Husain SZ, Bai H et al. (2012) Definitions of pediatric pancreatitis and survey of present clinical practices. J Pediatr Gastroenterol Nutr 55:261–265 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sun H, Zuo HD, Lin Q et al. (2019) MR imaging for acute pancreatitis: the current status of clinical applications. Ann Transl Med 7:269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Elhady M, Elazab AAAM, Bahagat KA et al. (2019) Fatty pancreas in relation to insulin resistance and metabolic syndrome in children with obesity. J Pediatr Endocrinol Metab 32:19–26 [DOI] [PubMed] [Google Scholar]
- 18.Staaf J, Labmayr V, Paulmichl K et al. (2017) Pancreatic fat is associated with metabolic syndrome and visceral fat but not beta-cell function or body mass index in pediatric obesity. Pancreas 46:358–365 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lee MS, Lee JS, Kim BS et al. (2021) Quantitative analysis of pancreatic fat in children with obesity using magnetic resonance imaging and ultrasonography. Pediatr Gastroenterol Hepatol Nutr 24:555–563 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Covarrubias Y, Fowler KJ, Mamidipalli A et al. (2019) Pilot study on longitudinal change in pancreatic proton density fat fraction during a weight-loss surgery program in adults with obesity. J Magn Reson Imaging 50:1092–1102 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang X, Misawa R, Zielinski MC et al. (2013) Regional differences in islet distribution in the human pancreas–preferential beta-cell loss in the head region in patients with type 2 diabetes. PLoS ONE 8:e67454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Savari O, Zielinski MC, Wang X et al. (2013) Distinct function of the head region of human pancreas in the pathogenesis of diabetes. Islets 5:226–228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chiyanika C, Chan DFY, Hui SCN et al. (2020) The relationship between pancreas steatosis and the risk of metabolic syndrome and insulin resistance in Chinese adolescents with concurrent obesity and non-alcoholic fatty liver disease. Pediatr Obes 15:e12653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wicklow BA, Griffith AT, Dumontet JN et al. (2015) Pancreatic lipid content is not associated with beta cell dysfunction in youth-onset type 2 diabetes. Can J Diabetes 39:398–404 [DOI] [PubMed] [Google Scholar]
- 25.Pimentel JL, Vander Wyst KB, Soltero EG et al. (2022) Organ fat in Latino youth at risk for type 2 diabetes. Pediatr Diabetes 23:286–290 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Toledo-Corral CM, Alderete TL, Hu HH et al. (2013) Ectopic fat deposition in prediabetic overweight and obese minority adolescents. J Clin Endocrinol Metab 98:1115–1121 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fujii M, Ohno Y, Yamada M et al. (2019) Impact of fatty pancreas and lifestyle on the development of subclinical chronic pancreatitis in healthy people undergoing a medical checkup. Environ Health Prev Med 24:10. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, S.E.S., upon reasonable request.