Parental Obesity and Offspring Serum Alanine and Aspartate Aminotransferase Levels: The Framingham Heart Study

Serum ALT and AST Levels by Parental History of Obesity 

Age- and sex-adjusted log ALT levels were higher among participants with at least one parent with early-onset obesity (Table 2) as compared with those without parental obesity (3.21 ± 0.04 vs 3.08 ± 0.02; P = .001). Age- and sex-adjusted log ALT levels remained statistically significant when participants with early-onset maternal or paternal obesity were compared with those without paternal or maternal obesity. The multivariable adjusted log ALT levels were significantly higher only among participants with paternal obesity as compared with those without paternal obesity (3.21 ± 0.06 vs 3.08 ± 0.01; P = .02). Similar comparison with respective referent groups for maternal and at least one parent obese models did not achieve statistical significance. No significant differences were observed in any of the models when participants with later-onset obesity were compared with those without parental obesity. No significant differences in AST levels in any of the models in the 3 exposure categories were observed.

Table 2. Log-Transformed ALT and AST Values by Parental History of Obesity

		Early-onset parental obesity (n = 193)	Later-onset parental obesity (n = 460)	No parental obesity (referent group) (n = 1079)	P valuea	P valueb
	ALT
	Father
	Age-sex model	3.27(0.06)	3.10(0.03)	3.08(0.02)	.001	.44
	Multivariable modelc	3.21(0.06)	3.08(0.03)	3.08(0.01)	.02	.99
	Mother
	Age-sex model	3.19(0.04)	3.07(0.03)	3.08(0.02)	.02	.72
	Multivariable modelc	3.15(0.04)	3.05(0.03)	3.10(0.02)	.22	.14
	At least one affected parent
	Age-sex model	3.21(0.04)	3.06(0.03)	3.08(0.02)	.001	.39
	Multivariable modelc	3.17(0.03)	3.03(0.03)	3.10(0.02)	.07	.05
	AST
	Father
	Age-sex model	3.00(0.05)	2.93(0.03)	2.91(0.01)	.06	.41
	Multivariable modelc	2.99(0.05)	2.92(0.03)	2.91(0.01)	.10	.63
	Mother
	Age-sex model	2.94(0.04)	2.92(0.03)	2.91(0.01)	.51	.70
	Multivariable modelc	2.93(0.04)	2.91(0.02)	2.90(0.01)	.69	.83
	At least one affected parent
	Age-sex model	2.96(0.03)	2.91(0.02)	2.90(0.02)	.15	.72
	Multivariable modelc	2.95(0.03)	2.90(0.02)	2.91(0.02)	.30	.77

NOTE. Least squares of mean (SE) of log-transformed serum ALT and AST values are presented. Results obtained from GEE are adjusted for familial correlation.

a Early-onset parental obesity versus no parental obesity. b Later-onset parental obesity versus no parental obesity. c Age, sex, diabetes, BMI, systolic blood pressure, hypertension treatment, total cholesterol/HDL ratio, triglyceride level, smoking status, total alcohol consumption, and cardiovascular disease.

Age- and sex-adjusted odds ratio (OR) of elevated serum ALT level was higher in offspring with both paternal early-onset obesity as compared with no paternal obesity (Table 3; OR, 2.04, 95% confidence interval [CI], 1.29–3.24, P = .005) and early-onset obesity in at least one parent compared with no parental obesity (OR, 1.52; 95% CI, 1.13–2.06; P = .006). However, statistical significance was not observed in models examining maternal early-onset obesity as compared with no maternal obesity (OR, 1.33; 95% CI, 0.95–1.85; P = .09). In multivariable models, participants with paternal early-onset obesity had higher serum ALT levels (Table 3; OR, 1.75; 95% CI, 1.06–2.89; P = .03) relative to those without paternal obesity. Early-onset paternal obesity was associated with ALT elevation independent of maternal obesity in multivariable-adjusted models (OR, 1.72; 95% CI, 1.02–2.69; P = .03). Those with a history of maternal obesity or at least one obese parent did not have elevated ALT levels, with an OR of 1.10 (95% CI, 0.76–1.59; P = .61) and an OR of 1.25 (95% CI, 0.90–1.74; P = .18), respectively, when compared with offspring without maternal obesity or those without either parent with obesity, respectively. There were no significant differences in the odds of elevated ALT levels in participants with later-onset parental obesity when compared with offspring without parental obesity (Table 3). The odds of elevated serum AST levels in the offspring did not differ in the exposure groups (early-onset parental obesity and later-onset parental obesity) when compared with the offspring without parental obesity.

Table 3. OR of Elevated ALT and AST Levels With or Without Adjustment Grouped by Parental Obesity Status

		Early-onset versus no parental obesity		Later-onset versus no parental obesity
		OR (95% CI)	P valuea	OR (95% CI)	P valueb
	ALT
	Father
	Age-sex model	2.04(1.29–3.24)	.005	1.05(0.79–1.39)	.76
	Multivariable modelc	1.75(1.06–2.89)	.03	0.93(0.68–1.27)	.66
	Mother
	Age-sex model	1.33(0.95–1.85)	.09	1.05(0.82–1.35)	.68
	Multivariable modelc	1.10(0.76–1.59)	.61	0.91(0.70–1.18)	.49
	At least one affected parent
	Age-sex model	1.52(1.13–2.06)	.006	1.01(0.80–1.26)	.97
	Multivariable modelc	1.25(0.90–1.74)	.18	0.88(0.69–1.12)	.29
	AST
	Father
	Age-sex model	1.26(0.78–2.04)	.34	0.94(0.69–1.29)	.71
	Multivariable modelc	1.18(0.73–1.91)	.49	0.91(0.65–1.26)	.56
	Mother
	Age-sex model	1.24(0.81–1.74)	.28	0.95(0.72–1.26)	.73
	Multivariable modelc	1.14(0.75–1.71)	.54	0.86(0.65–1.16)	.33
	At least one affected parent
	Age-sex model	1.28(0.92–1.80)	.15	0.95(0.73–1.23)	.68
	Multivariable modelc	1.17(0.82–1.66)	.39	0.87(0.66–1.14)	.32

NOTE. The analysis is based on the GEE model adjusted for familial correlation. The referent group is either offspring without a paternal, maternal, or parental history of obesity for father, mother, and at least one affected parent model, respectively.

a Early onset versus no parental obesity. b Later-onset versus no parental obesity. c Age, sex, diabetes, BMI, systolic blood pressure, hypertension treatment, total cholesterol/HDL ratio, triglyceride level, smoking status, total alcohol consumption (drinks/week), and cardiovascular disease.

Secondary Analyses 

After excluding obese offspring, the overall findings were essentially unchanged, although in these secondary analyses the multivariable adjusted ORs were somewhat strengthened for the association between paternal early-onset obesity and ALT (see Supplementary Table 1; see supplemental material online at www.gastrojournal.org). Among individuals without excess alcohol consumption or abuse, the results were not substantially different (data not shown).

Principal Findings 

In this community-based sample of white adults, we found that a paternal history of early-onset obesity is associated with a higher OR of elevated serum ALT levels in offspring. This relation is independent of the offspring’s BMI and persists among nonobese offspring. Moreover, early-onset paternal obesity but not maternal obesity was associated with elevated serum ALT levels in the offspring. On the contrary, serum AST levels were not associated with parental obesity status.

In the Context of Current Literature 

NAFLD is now considered the most common cause of serum ALT elevations in the US population.9 Previous studies have shown that there is familial clustering of factors that may predispose to the development of NAFLD, NASH, and cryptogenic cirrhosis.18, 19, 26 Struben et al reported increased prevalence of NASH or cryptogenic cirrhosis or both in 7 of 8 kindreds.18 Willner et al showed that the prevalence of NAFLD is higher in first-degree relatives of patients with NASH.19 Both of these findings suggest that NAFLD and elevated serum ALT levels may have an underlying genetic susceptibility, particularly in the setting of obesity.27, 28 A recent study by Kazumi et al showed an association between parental BMI and elevated serum ALT levels in their male offspring.29 Contrary to our findings, this study reported that maternal rather than paternal BMI was associated with elevated ALT levels in Japanese men. The primary difference between this prior study and ours, in addition to being conducted in a sample of men from Japan, classified parental BMI exposure based on offspring recall. In our study, detailed information regarding measurement of parental BMI during a physician visit was obtained. Previous studies have shown that patient recall can lead to misclassification of exposure status. Therefore, differences in age, sex, race, and parental BMI may contribute to the disparate findings between these studies.

Mechanism 

Our findings suggest that there may be a familial component to ALT levels, and in particular, this predisposition may be mediated through early-onset familial obesity. Obesity has a strong familial and genetic component.14 Early-onset or premature occurrence of a disease condition may be more associated with an underlying genetic susceptibility. Therefore, these data support a potential mechanism of elevated ALT level.

Several genes have been implicated in the development of early-onset obesity. A missense mutation in human pro-opiomelanocortin (POMC) gene30 and single nucleotide polymorphisms in melanin-concentrating hormone receptor 1 (MCHR1) have been associated with development of early obesity.31 It is known that mutations in the leptin gene lead to a severe form of early-onset obesity.32 Leptin is believed to be an important regulator of fat and energy metabolism,33 and several studies have shown that leptin may play a key role in regulating hepatic fibrosis34, 35 and progression of NASH.36 Our findings raise the possibility that genes involved in the pathogenesis of early-onset obesity may also be associated with abnormal ALT levels. However, we cannot rule out the possibility that shared environmental and unmeasured behavioral practices could have contributed to this finding.

Our findings were also notable for a stronger effect in the presence of a paternal history of early-onset obesity. One potential explanation may be that serum ALT is linked to genes on the Y-chromosome or X-linked conditions. An additional explanation may have to do with gene imprinting or telomere length, both of which have been suggested as a mechanism of paternal mode of inheritance.37, 38 However, these hypotheses are speculative and need to be explored in future studies.

Our results showed association for ALT and not AST. This was not entirely surprising because ALT has been shown to correlate better with obesity.39 Furthermore, elevations of serum AST level are more sensitive to alcohol consumption as compared with ALT, which is usually associated with NAFLD.40 Lastly, serum ALT is a more specific marker of liver injury than serum AST.41, 42

Strengths and Limitations 

Strengths of our study include the use of a well-characterized large community-based sample. Parental adiposity status was obtained from routine Framingham Heart Study clinic examinations and was not based on offspring self-report. We were able to account for potentially important confounding variables that are known to be related to elevated serum ALT levels. Our study has several limitations. First, our sample was exclusively white; thus, the generalizability of these findings to other races or ethnic groups is uncertain. Second, the study sample included only offspring with both parents in the Framingham Heart Study. This was necessary to minimize bias due to misclassification of parental obesity status. Third, parental serum ALT and AST levels were not measured. Therefore, we cannot exclude the possibility that parents with early-onset obesity also had abnormal ALT or AST levels. Information on other liver diseases, such as viral hepatitis or autoimmune or metabolic liver disease and their risk factors, was not available, and therefore these conditions could not be excluded as a cause of ALT elevations. To limit potential bias, we excluded individuals who might have these conditions by restricting our study sample to offspring with ALT and AST levels <120 IU/L. These individuals were excluded because higher levels of serum ALT and AST are more likely to be associated with acute liver injury that may be related to factors such as drug-induced liver injury and acute viral hepatitis and not NAFLD. Nonetheless, inclusion of these individuals in our study would result in misclassification and would bias the results toward the null hypothesis. Therefore, it is unlikely that this issue contributes to our findings. Lastly, we did not have information on central obesity or measures of insulin resistance, 2 factors that could potentially mediate our findings.

Implications for Further Research 

Our findings support the hypothesis that familial factors may play a role in the risks of elevated serum ALT levels in the general population. Further, this association may be mediated through early-onset obesity. It is possible that genes that associate with early-onset obesity may also be associated with elevated ALT levels. Future studies should specifically examine the relations of potential candidate genes, elevated ALT levels, and NAFLD. However, it is important to recognize that our findings do not establish a causal relationship between genetic factors and the development of elevated serum ALT levels or NAFLD.

These results support the need for further studies to establish whether individuals with early-onset parental obesity and elevated serum ALT levels are at a higher risk for developing progressive liver disease such as NASH. Further research will also be important to identify whether these individuals are also at increased risk for developing metabolic complications of both obesity and NAFLD. These results need to be validated in other family studies to further understand the role of genetic and shared environmental factors on susceptibility to developing elevated serum ALT levels.