NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Guideline Centre (UK). Non-Alcoholic Fatty Liver Disease: Assessment and Management. London: National Institute for Health and Care Excellence (UK); 2016 Jul. (NICE Guideline, No. 49.)

Cover of Non-Alcoholic Fatty Liver Disease

Non-Alcoholic Fatty Liver Disease: Assessment and Management.

Show details

5Risk factors for NAFLD

5.1. Introduction

Non-alcoholic fatty liver disease (NAFLD) – ranging from simple fatty liver, through non-alcoholic steatohepatitis (NASH) and fibrosis to NASH-cirrhosis – is highly prevalent, being present in more than 20% of the general population in Europe and North America and higher in the Middle East and South Asia. In most studies, the prevalence is higher in individuals with features of metabolic syndrome including obesity, diabetes, hypertension and an increased waist circumference. Furthermore, several studies have suggested that the presence of these factors also increases the likelihood that people have more advanced forms of NAFLD (fibrosis and NASH). In light of these reports, this review investigates which risk factors (or combination of risk factors) are good indicators of NAFLD and NASH or fibrosis.

5.2. Review question: Which risk factors for NAFLD or severe NAFLD (NASH, fibrosis) aid in the identification of people who should be investigated further?

For full details see review protocol in Appendix C.

Table 8. Characteristics of review question.

Table 8

Characteristics of review question.

5.3. Clinical evidence

Six studies were included in the review altogether, all of which were based on an adult population.62,84,96,180,185,219 Evidence from these are summarised in the clinical GRADE evidence profile below. See also the study selection flow chart in Appendix E, forest plots in Appendix K, study evidence tables in Appendix H and exclusion list in Appendix M.

All 6 of the adult studies were cohort studies with a multivariate analysis model which adjusted for 3 or more confounders outlined in the protocol, investigating the association of the prognostic variables and NAFLD. There were no cohort studies identified investigating the association of the prognostic variables with NASH or fibrosis in adults. There were also no cohort studies identified investigating the association between the prognostic factors and NAFLD, NASH or fibrosis in young people and children.

Table 9. Summary of studies included in the review.

Table 9

Summary of studies included in the review.

5.3.1. Risk factors for NAFLD

Table 10. Waist circumference as a risk factor for NAFLD (adults).

Table 10

Waist circumference as a risk factor for NAFLD (adults).

Table 11. Hypertension as a risk factor for NAFLD (adults).

Table 11

Hypertension as a risk factor for NAFLD (adults).

Table 12. Triglycerides as a risk factor for NAFLD (adults).

Table 12

Triglycerides as a risk factor for NAFLD (adults).

Table 13. LOW HDL-cholesterol as a risk factor for NAFLD (adults).

Table 13

LOW HDL-cholesterol as a risk factor for NAFLD (adults).

Table 14. Type 2 diabetes as a risk factor for NAFLD (adults).

Table 14

Type 2 diabetes as a risk factor for NAFLD (adults).

Table 15. Age as a risk factor for NAFLD (adults).

Table 15

Age as a risk factor for NAFLD (adults).

Table 16. BMI as a risk factor for NAFLD (adults).

Table 16

BMI as a risk factor for NAFLD (adults).

5.3.1.1. Metabolic syndrome

Table 17. Risk factor (metabolic syndrome) for NAFLD (adults).

Table 17

Risk factor (metabolic syndrome) for NAFLD (adults).

5.3.2. Risk factors for NASH

No clinical evidence was identified.

5.3.3. Risk factors for fibrosis

No clinical evidence was identified.

5.4. Economic evidence

5.4.1. Published evidence

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix F.

5.4.2. New cost-effectiveness analysis

Original cost-effectiveness modelling was undertaken for this question using the NGC liver disease pathway model developed for this guideline. A summary of the modelling work and an evidence statement summarising the results of the analysis can be found in Chapter 6. The full analysis can be found in Appendix N.

The following risk factors were considered in the analysis:

  • high triglycerides (≥1.5 g/litre)
  • low HDL-cholesterol (<400 mg/litre for men, <500 mg/litre for women)
  • metabolic syndrome (NCEP criteria)
  • obesity (BMI≥30)
  • type 2 diabetes (glycaemia≥1.1 g/litre)
  • wide waist circumference (≥102 cm for men, ≥88 cm for women)

5.5. Evidence statements

5.5.1. Clinical

5.5.1.1. Prognostic variables as risk factors for NAFLD

Waist circumference
  • High quality evidence from 2 cohort studies with multivariable analysis suggested waist circumference to be a predictor of NAFLD in adults, with an adjusted HR of 1.08 (1.06, 1.10; n=6905) and an adjusted OR of 1.08 (1.06, 1.10; n=3577).
  • There was no evidence for children and young people.
Hypertension
  • Moderate to low quality evidence from 4 cohort studies, with multivariable analysis, showed hypertension and blood pressure were not significant predictors of NAFLD in a general, obese and non-obese adult population. Adjusted HRs in one of the studies (n=1705) was 0.99 (0.72, 1.36) for BP (general population), and in the other (n=6905) was 1.01 (1.0, 1.02) and 1 (0.99, 1.01), diastolic BP (non-obese population) and systolic BP (non-obese population) respectively. Adjusted ORs reported were 0.90 (0.64, 1.27) and 1.18 (0.93, 1.48) for hypertension (general population; n=2029) and BP (mixed obese and non-obese population; n=2307) respectively
  • There was no evidence for children and young people.
Triglycerides
  • There was some evidence suggesting that adults with high triglycerides may be at higher risk of NAFLD from 5 cohort studies with multivariable analysis. High quality evidence from 1 study (n=1705) reported an adjusted HR of 2.10 (1.52, 2.90) and high to moderate quality evidence from 4 prospective cohort studies (with sample sizes ranging from 2307 to 6905) reported an adjusted HR of 1.21 (1.07, 1.37) and adjusted ORs of 1.38 (1.18, 1.61), 1.25 (1.19, 1.31) and 1.41 (1.11, 1.8).
  • There was no evidence for children and young people.
HDL-cholesterol
  • There was some evidence suggesting that adults with lower HDL-cholesterol may be at higher risk of NAFLD from 2 cohort studies with multivariable analysis. High quality evidence from 1 study (n=6905) suggested that low HDL cholesterol was predictive of NAFLD with an adjusted HR of 0.57 (0.34, 0.96), along with low quality evidence from another study which reported an adjusted OR of 0.82 (0.55, 1.24). However, low quality evidence from a second study (n=1705) suggested low HDL was protective of NAFLD with an adjusted HR of 1.23 (0.91, 1.66).
  • There was no evidence for children and young people.
Diabetes
  • Moderate quality evidence from 1 study (n=3529), with multivariable analysis, suggested type 2 diabetes as a risk factor for NAFLD reporting an adjusted OR of 1.64 (1.11, 2.42).
  • There was no evidence for children and young people.
Age
  • High to moderate quality evidence from 4 cohort studies, with multivariable analysis, showed age was not a significant predictor of NAFLD in adults. Three studies (with sample sizes ranging from 2029 to 3577) reported adjusted ORs of 0.95 (0.94, 0.97), 1.03 (1.02, 1.04) and 0.99 (0.98, 1.0); and 1 study (n=6905) reported adjusted HR of 0.98 (0.97, 0.99).
  • There was no evidence for children and young people.
BMI
  • High to moderate quality evidence from 3 cohort studies (sample sizes ranging from 1705 to 6905), with multivariable analysis, suggests some evidence that adults with a high BMI may be at risk of NAFLD reporting adjusted HRs of 2.46 (1.88, 3.22) and 1.22 (1.13, 1.32); and an adjusted OR of 1.14 (1.04, 1.26).

There was no evidence for children and young people.

Metabolic syndrome (combination of risk factors)
  • There was some high to moderate quality evidence suggesting that adults with metabolic syndrome may be at higher risk of NAFLD from 3 cohort studies with multivariable analysis (sample sizes ranging from 1705 to 3529). Two high quality studies reported adjusted ORs of 1.95 (1.48, 2.57) and 2.99 (1.62, 5.52) and 1 study reported a HR of 5.91 (3.93, 8.89).

There was no evidence for children and young people.

5.5.1.2. Prognostic variables as risk factors for NASH or fibrosis

There was no evidence in adults or children and young people to determine the risk factors for NASH or fibrosis.

5.5.2. Economic

  • One original cost-utility analysis found that testing for NAFLD was cost-effective compared to no testing at a cost-effectiveness threshold of £20,000 per QALY gained for all retest frequencies and NAFLD prevalence's investigated. Retesting at a frequency of 6 years was cost-effective compared to other frequencies. This analysis was assessed as directly applicable with minor limitations.

5.6. Recommendations and link to evidence

Recommendation
  1. Be aware that non-alcoholic fatty liver disease (NAFLD) is more common in people who have:
    • type 2 diabetes or
    • metabolic syndrome.
Research recommendation
  1. What are the prognostic factors for the development of NAFLD or NASH in children and young people?
Relative values of different prognostic measures and outcomesThe GDG agreed that diagnosis of NAFLD and diagnosis of NASH and fibrosis were the most informative outcomes for this review. However, no cohort study evidence was identified for the outcome diagnosis of NASH and fibrosis.
Trade-off between clinical benefits and harmsThe GDG concluded that the evidence identified within this part of the review demonstrated that waist circumference, type 2 diabetes mellitus, increased triglycerides, low HDL-cholesterol and metabolic syndrome appeared to be predictive factors for NAFLD in adults; with metabolic syndrome being the single most strongly predictive factor. It was noted that metabolic syndrome seemed to be predictive of NAFLD in adults regardless of whether people had all components of the syndrome or only certain elements of it (irrespective of which combination of elements they were). The GDG noted that, whilst there was some association between increased BMI in adults and NAFLD, the strength of this association varied significantly between studies. There was no compelling evidence that hypertension or age were predictive of NAFLD in adults. Furthermore, for all the risk factors considered, there was little evidence to suggest a dose-response relationship with the development of NAFLD.
The GDG decided that the risk factors that merited inclusion in the economic modelling as predictive factors for NAFLD for adults were type 2 diabetes mellitus, high triglycerides, high BMI, low HDL-cholesterol, wide waist circumference and the presence of metabolic syndrome. The GDG concluded that the lack of any cohort study evidence for risk factors in children and young people meant that it would not be possible to model risks in this population; therefore only adults were considered in the economic model. However, the GDG agreed that there was no specific reason to suggest that these risk factors would differ in a younger population and agreed to extrapolate the evidence from adult populations to make a recommendation for all age groups. A research recommendation for further research on risk factors for NAFLD in children and young people was considered appropriate to confirm this assumption.
This review also looked for predictive factors for NASH and severe fibrosis. This is because a major aim of both primary and secondary care is to identify the smaller number of ‘high risk’ people with NAFLD who merit further intervention (because they also have NASH or significant fibrosis) from the large numbers of people within the general population with ‘simple’ steatosis. However, no evidence was identified on predictive factors for these conditions and the GDG agreed that it was unable to make a recommendation on whether the presence of any specific risk factors might warrant testing for NASH or severe fibrosis.
Trade-off between net clinical effects and costsNo relevant published economic evaluations were identified.
An original cost-utility analysis was conducted for this guideline to address the questions in this review and also Chapters 6, 7 and 8. As described above, no economic analysis was conducted relating to children and young people under 18 years.
The risk factor groups with the highest prevalence of NAFLD were people with metabolic syndrome (54%) and people with type 2 diabetes (53%). As explained in further detail in Chapter 6, the original economic analysis compared the cost-effectiveness of 8 diagnostic tests with 2 no testing strategies, and found the fatty liver index (FLI) was cost-effective compared to all of the other diagnostic tests. The analysis found that FLI was also cost-effective compared to no testing (with no treatment) for adults with type 2 diabetes or metabolic syndrome at all frequencies of retesting investigated (1–8 years) at a cost-effectiveness threshold of £20,000 per QALY gained.
For adults with a high BMI (46% prevalence of NAFLD), high triglycerides (46%), low HDL-cholesterol (36%), or wide waist circumference (36%), testing using FLI was also cost-effective compared to no testing at all retesting frequencies investigated. However, given that routine testing for NAFLD will be a new activity, and conscious of the number of people involved, the GDG agreed that testing for NAFLD should be prioritised to those groups (people with type 2 diabetes or metabolic syndrome) at highest risk of having or developing NAFLD. The GDG further noted that people with any of the other individual risk factors would themselves be at risk of developing metabolic syndrome or type 2 diabetes over the coming years. At the point that they do develop one of those conditions, they would then become eligible for NAFLD testing.
See Chapter 6 for recommendations on the test to be used to diagnose NAFLD.
Quality of evidenceThe clinical studies included in this review were cohort studies based on adult populations that adjusted for 3 or more confounders listed in the review protocol. The quality of the majority of these study outcomes ranged from a modified GRADE rating of moderate to low. Downgrading of evidence was predominantly due to the high risk of bias in some studies due to the lack blinding, as well as the imprecise nature of the results. Evidence of a high GRADE rating was seen for some of the evidence included; for the prognostic factors waist circumference, triglycerides, low HDL-cholesterol, age, BMI and metabolic syndrome.
The economic study included in this review was of high quality, being directly applicable and with minor limitations.
Other considerationsResearch recommendation
The GDG made a research recommendation to elucidate the prognostic factors for NAFLD and NASH in children and young people due to the lack of published studies in this area.
Copyright © National Institute for Health and Care Excellence 2016.
Bookshelf ID: NBK384735

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...