Prevalence of Risk Factors for Nonalcoholic Fatty Liver Disease in Middle-Aged and Elderly Patients With Cryptogenic Cirrhosis

Abstract

Aim of the study

To study the prevalence of risk factors for nonalcoholic fatty liver disease (NAFLD) in middle-aged (40–59 years) and elderly patients (≥60 years) with cryptogenic cirrhosis as compared to those with hepatitis B or C virus (HBV or HCV) related cirrhosis.

Methods and materials

Between August 2013 and December 2014, cases (cryptogenic cirrhosis) and controls (HBV/HCV cirrhosis) above 40 years of age were prospectively recruited and assessed for the cause and prevalence of risk factors for NAFLD.

Results

One hundred eighteen cases (male-74%; age 55 (40–74) years; median (range); Child's class A:B:C-46:38:16) and 59 controls (male-80%; age 55.5 (40–69) years; Child's class A:B:C-56:30:14) were enrolled. Obesity (53% v/s 39%, P-0.081), diabetes mellitus (DM) (52% v/s 27%; P-0.002), family history of DM (30% v/s 13%; P-0.016), family history of Obesity (21% v/s 3.5%; P-0.002) and metabolic syndrome (65% v/s 44%; P-0.01) were more among cases than controls. Lifetime weight as obese was also longer in cases than in controls (5.9 ± 6.2 years v/s 3.2 ± 5.1 years, P-0.002). On subgroup analysis, in elderly age group, DM (55% v/s 17%, P-0.006), family history of DM (40% v/s 11%, P-0.025), metabolic syndrome (76% v/s 44%, P-0.017) and family history of obesity (19% v/s 0, P-0.047) were more common in cases as compared to controls, where as in the middle-age group, family history of obesity was the only significant factor (22% v/s 5%, P-0.025). Lifetime weight as obese was longer in cases than controls in both middle and elderly age groups.

Conclusion

Among middle-aged and elderly patients with cirrhosis, there was a higher prevalence of risk factors for NAFLD in those with cryptogenic cirrhosis, compared to those with HBV or HCV cirrhosis.

The word “cryptogenic” meaning secret, something hidden, originates from Greek words “kryptos” and “genesis”, meaning hidden and origin, respectively. Cryptogenic cirrhosis is diagnosed if the aetiological work up is negative for any known etiological factor such as hepatitis B virus (HBV), hepatitis C virus (HCV), alcohol, haemochromatosis, Wilsons disease or autoimmune hepatitis.

Patients with cryptogenic cirrhosis have an increased prevalence of metabolic syndrome and its components when compared to controls. Metabolic syndrome and its components are recognised as an important risk factor for cryptogenic cirrhosis.¹ Studies from India have shown that, in cases of unexplained elevation of liver enzymes, cryptogenic cirrhosis or cryptogenic hepatocellular carcinoma, NAFLD is an important etiological factor.² NAFLD is seen in 9%–12% of the general population and have higher prevalence of obesity, overweight, prediabetes, diabetes and metabolic syndrome, majority of these patients were also noted to have insulin resistance.² Nayak et al in their study on explant livers reported that viral hepatitis-related cirrhosis was the most common cause for end-stage cirrhosis, accounting to 48.6%, whereas 23.1% was alcohol related and 16.7% was NAFLD related.³ Poonawala et al noted a higher prevalence of obesity and type II diabetes mellitus in patients with cryptogenic cirrhosis compared with cirrhosis of other causes.⁴

Studies from around the world have accounted NAFLD as one of the major etiological factors of cryptogenic cirrhosis.3, 4, 5, 6, 7 If appropriate criteria are applied to cases labelled as cryptogenic cirrhosis, NAFLD can account for almost two-thirds of them.⁸ In a study by Charlton et al, patients with cryptogenic cirrhosis had higher Child–Pugh score and worse survival after liver transplantation, as compared to patients with cirrhosis of other aetiologies.⁹

Cryptogenic cirrhosis remains one of the important causes of cirrhosis and NAFLD is one of the aetiologies of cryptogenic cirrhosis. As cirrhosis progresses, diagnosis of NAFLD is difficult and even on liver biopsy, there may be no evidence of steatosis. Furthermore, the prevalence of obesity, diabetes mellitus and hypertension, which are the risk factors for NAFLD shows a rising trend in the country. Also as the age advances the risk of developing these risk factors increases. There are limited studies from India on cryptogenic cirrhosis and the prevalence of risk factors for NAFLD. This study was planned with the objective of estimating the preva-lence of risk factors for NAFLD in middle-aged and elderly patients with cryptogenic cirrhosis as compared to HBV- or HCV-related cirrhosis.

Methodology

This study was a prospective, case-control study analysing the prevalence of risk factors of NAFLD in patients with cryptogenic cirrhosis with an emphasis on middle aged and elderly. The study was conducted between August 2013 to December 2014 in the Department of Gastroenterology and Hepatology, CMC Vellore, a tertiary care centre in southern India. Consecutive (new) patients presenting with cryptogenic cirrhosis were recruited as cases, and patients with either HBV- or HCV-related cirrhosis served as controls.

Participants

Cases

Consecutive patients above 40 years of age presenting with cryptogenic cirrhosis (ultrasound scan showing coarse echotexture with volume redistribution with or without splenomegaly, or evidence of portal hypertension on upper GI scopy, or cirrhosis diagnosed by biopsy and no known cause of cirrhosis) were categorised as cases. Patients with cirrhosis of known aetiologies such as alcohol, viral hepatitis, haemochromatosis, autoimmune hepatitis, Wilson disease, drug induced liver disease or primary biliary cirrhosis were excluded. Patients with evidence of cirrhosis or portal hypertension in the absence of a known aetiological factor after work up were considered to have cryptogenic cirrhosis.

Controls

Consecutive patients above 40 years of age having HBV- or HCV-related cirrhosis were recruited as controls.

In both arms, patients with significant alcohol intake (men >60 g per day and women >20 g per day for 10 years or more) were excluded. Those who did not consent for the study were excluded. Study subjects 60 years and above were considered as elderly and those between 40 and 59 years as middle aged.

Variables Measured in Patients

• 1)
  Risk factors for NAFLD

  • i)
    Obesity: We used suggested cutoffs of body mass index (BMI) for Asian-Indians to classify obesity (normal-18 to 22.9 kg/m², overweight-23 to 24.9 kg/m² and obesity->25 kg/m².¹⁰
  • ii)
    Diabetes mellitus: Patients were diagnosed to have diabetes mellitus with either fasting glucose > 126 mg% or 2-h postprandial glucose >200 mg%. Patients, who were detected to be diabetic elsewhere and were on oral hypoglycemic agents/insulin at presentation here, were considered to be diabetic even if they had normal glucose values on testing.
  • iii)
    Waist hip ratio (WHR): waist-hip ratio above 0.90 in men and 0.85 in women were taken as abnormal. Waist circumference greater than 90 cm [men] and 80 cm [women] as described for Indian population was considered as elevated.¹⁰
  • iv)
    Dyslipidemia: Dyslipidemia was defined as either serum triglycerides ≥150 mg/d or total cholesterol ≥ 200 mg% or LDL if ≥130 mg% and also those who were on lipid lowering medications.¹¹
  • v)
    Hypertension: Blood pressure higher than 140/90 mmHg on two separate occasions or those who were on antihypertensive medications.¹²
  • vi)
    Family history of diabetes mellitus: Any first degree relative having diabetes mellitus.¹³
  • vii)
    Lifetime body weight: was defined as the number of years of overweight or obese.
  • viii)
    Metabolic syndrome: Metabolic syndrome was diagnosed by the presence of at least three of the following criteria, in accordance with the ATP III guidelines (modified for Asian Indians).¹⁰

    • a)
      Elevated waist circumference – Males >90 cm, females >80 cm
    • b)
      Elevated triglycerides – ≥150 mg/dl
    • c)
      Reduced HDL – <40 mg/dl (1.0 mmol/L) in males; <50 mg/dl (1.3 mmol/L) in females
    • d)
      Elevated blood pressure – Systolic ≥130 and/or diastolic ≥85 mm Hg
    • e)
      Elevated fasting glucose – >100 mg/dl
  • ix)
    Premorbid body weight: Weight before the diagnosis of cirrhosis.
  • x)
    Family history of hypertension: Any first degree relative having hypertension.
  • xi)
    Family history of dyslipidemia: Any first degree relative having dyslipidemia.
  • xii)
    Family history of obesity: Any first degree relative having obesity.

Family history of dyslipidemia, obesity, hypertension and DM were based on the history given by the patients.

• 2)Occult hepatitis B infection – All patients with cryptogenic cirrhosis (cases) who were HBsAg negative were tested for total anti hepatitis B core antibody to look for occult HBV infection. Patients who were total anti hepatitis B core antibody positive were advised HBV DNA viral load to confirm hepatitis B infection.

Statistical Methods

Sample Size Calculation

Sample size was calculated using software N-master version 2. With an estimated prevalence of 40% of risk factors for NAFLD in cryptogenic cirrhosis a sample size of 118 cases and 59 controls was required to detect a difference of proportion of 20% with 80% power and 5% alpha error. Age-wise subgroup analysis was done subsequently.

Statistical Analysis

The statistical analysis was performed using SPSS version 17. Continuous variables were expressed as median (range) and/or mean ± standard deviation and number and/or percentage was used to express categorical variables. Chi-square test was used to test the differences between the categorical parameters, whereas the difference between groups for continuous variables was calculated using independent t-tests. Odds ratio was calculated for the risk factors between cases and controls with 95% confidence interval to assess the risk of association. A P value of <0.05 was considered statistically significant. Multivariate analysis was performed to identify the independent risk factors. A P-value of <0.05 and confidence intervals (CIs) not including 1 were considered to be statistically significant.

The study was approved by the Institutional Review Board and Ethics committee.

Results

Patients with cryptogenic cirrhosis (n-118) were included as cases and patients with either hepatitis B or hepatitis C related cirrhosis (n-59) were included as controls (hepatitis B: 30; hepatitis C: 29). The patient flow is depicted in Figure 1.

Figure 1.

Patient flow chart.

Baseline variables of the patients including demographic characteristics, laboratory data and clinical features are presented in Table 1 (Figure 2). Majority of the cases and controls were males. Child's score distribution was similar in both cases and controls, with majority of cases and controls being Child's A status. Median MELD score was also similar in both cases and controls. Among cases (n-118), 61 patients were from West Bengal, 42 patients from Tamil Nadu, 7 patients from Kerala, 4 patients from Jharkhand, 3 patients from Bihar and 1 patient from Andhra Pradesh. Among controls (n – 59), 28 patients were from West Bengal, 22 patients from Tamil Nadu, 4 patients from Andhra Pradesh, 3 patients from Jharkhand and 2 patients from Bihar.

Table 1.

Baseline characteristics and clinical features.

Baseline characteristics and clinical features	Cases (n-118)	Controls (n-59)	P value
Age {in years, median (range)}	55 (40–74)	55.5 (40–69)	0.290
Sex {male, n (%)}	87 (74%)	47 (80%)	0.459
Bilirubin – Total {median (range)}	1.10 (0.2–10.10)	1.17 (0.3–39.10)	0.379
Albumin (mg/dl, mean ± SD)	3.4 ± 0.75	3.3 ± 0.9	0.117
Aspartate aminotransferase (mg/dl, mean ± SD)	52 ± 52	69 ± 65	0.022
Alanine aminotransferase (mg/dl, mean ± SD)	29 ± 21	39 ± 30	0.083
Alkaline phosphatase (IU, mean ± SD)	120 ± 64	118 ± 84	0.824
Creatinine (mean ± SD)	1.1 ± 0.5	1.01 ± 0.03	0.237
INR (mean ± SD)	1.2 ± 0.23	1.25 ± 0.2	0.902
MELD {median (range)}	10 (6–25)	10 (6–40)	0.757
CTP status {n (%)}
A	54 (46%)	33 (56%)	0.441
B	45 (38%)	18 (30%)
C	19 (16%)	08 (14%)
Complications {n (%)}
Evidence of portal hypertension	105 (89%)	46 (78%)	0.070
Ascites	63 (53%)	26 (44%)	0.338
UGI bleed	18 (15%)	05 (8.5%)	0.243
Hepatic encephalopathy	42 (36%)	15 (26%)	0.175
Acute kidney injury	07 (06%)	03 (05%)	1.000
Spontaneous bacterial peritonitis	11 (09%)	07 (12%)	0.605
Hepatocellular carcinoma	06 (05%)	02 (03%)	0.721

Figure 2.

Number of risk factors for Non-alcoholic fatty liver disease (NAFLD).

Risk Factors for NAFLD

Diabetes mellitus was more prevalent in cases (52%) as compared to controls (27%), (P value-0.002) and 30% of cases and 13% of controls had a family history of diabetes mellitus (P-0.016). The median duration of diabetes in cases compared to controls was not significant (Median/IQR – 7/0.5 years versus 5/0 years, P-0.113). Obesity as well as a family history of obesity were more common in cases than in controls. Lifetime weight as obese was longer in cases (5.9 ± 6.2 years) as compared to controls (3.2 ± 5.1 years, P-0.002). Metabolic syndrome was noted in 65% cases compared to 44% of controls (P-0.010).

Odds ratio was found to be statistically significant (P < 0.05) among cases with regard to diabetes mellitus, family history of diabetes mellitus, family history of obesity and metabolic syndrome (Table 2).

Table 2.

Risk Factors for Nonalcoholic Fatty Liver Disease (NAFLD).

	Cases (n-118)	Controls (n-59)	Odds ratio	95% confidence interval		P value
				Upper	Lower
Dyslipidemiaa {n (%)}	25 (21%)	11 (19%)	1.233	0.562	2.708	0.697
Family history of dyslipidemia {n (%)}	10 (09%)	08 (14%)	0.590	1.259	4.514	0.302
Diabetes mellitus {n (%)}	61 (52%)	16 (27%)	2.876	1.460	5.666	0.002
Family history of DM n {n (%)}	35 (30%)	08 (13%)	2.799	1.206	6.436	0.016
Obesity {n (%)}	63 (53%)	23 (39%)	1.793	0.949	3.387	0.081
Family history of obesity {n (%)}	25 (21%)	02 (3.5%)	7.661	1.748	33.57	0.002
Hypertension {n (%)}	32 (27%)	15 (25%)	1.091	0.535	2.226	0.858
Family history of hypertension {n (%)}	13 (11%)	04 (07%)	1.702	0.530	1.585	0.430
Metabolic syndrome {n (%)}	77 (65%)	26 (44%)	2.384	1.259	4.514	0.010
Total cholesterol (mg/dl, mean ± SD)	124 ± 43	113 ± 46	NA			0.494
Triglycerides (mg/dl, mean ± SD)	90 ± 38	90 ± 42	NA			0.938
High-density lipoprotein (mg/dl, mean ± SD)	36 ± 15	34 ± 18	NA			0.199
Low-density lipoprotein (mg/dl, mean ± SD)	72 ± 34	67 ± 34	NA			0.701
Glucose-fasting (mean ± SD) mg/dl	117 ± 33	113 ± 47	NA			0.168
Glucose-postprandial (mean ± SD) mg/dl	182 ± 76	170 ± 74	NA			0.572
Duration of DM in years{median (IQR)}	7 (0.5)	5 (0)	NA			0.113
Weight {kg, median (range)}	67.6 (40–100.8)	63.7 (45–90.1)	NA			0.059
Waist-hip ratio {median (range)}	1.02 (0.89–2.0)	1.02 (0.91–1.11)	NA			0.237
BMIa {kg/m2, median (range)}	25.2 (16.60–35.7)	23.6 (17.3–32.9)	NA			0.138
Lifetime weight as obese (years, mean ± SD)	5.9 ± 6.2	3.2 ± 5.1	NA			0.002
Premorbid weight (kg, mean ± SD)	70 ± 13	64 ± 10	NA			0.033
≥3 risk factors {n (%)}	33 (28%)	14 (24%)	NA			0.592

HCV, hepatitis C virus; RNA, ribonucleic acid.

^aDyslipidemia-elevated triglycerides >150 mg/dl or total cholesterol >200 mg/dl or LDL >130 mg/dl.

One or more risk factors for NAFLD was noted in 79% of cases and 73% of controls (P-0.150). ≥ 3 risk factors were seen in 28% cases and 24% controls (P-0.592). Risk factors found were diabetes mellitus, hypertension, dyslipidemia, obesity, metabolic syndrome and family history of diabetes mellitus. None of the patients among cases or controls had been previously diagnosed with NASH (nonalcoholic steatohepatitis) or NAFLD as the cause of their liver disease.

Subgroup analysis showed no difference in the baseline characteristics between middle aged and elderly. In the middle-aged group 22 controls had HBV related cirrhosis and 19 controls had HCV related cirrhosis. Among the elderly, 8 controls had HBV related cirrhosis and 10 controls had HCV related cirrhosis.

Age-wise Comparison of Risk Factors

Risk factors for NAFLD in both middle-aged and elderly subjects are shown in Table 3. In the middle-age group, lifetime weight as obese was longer in cases (5.5 ± 5.7 years) as compared to controls (3.3 ± 5.2years, P-0.045). Family history of obesity was noted in 22% of cases as compared to 5% (P-0.014). Among elderly subjects, diabetes mellitus (55% vs. 17%, P-0.006), family history of diabetes mellitus (40% vs. 11%, P-0.025), metabolic syndrome (76% vs. 44%, P-0.017) and family history of obesity (19% vs. 0, P-0.047) was more common in cases as compared to controls, whereas lifetime weight as obese was longer in cases (6.6 ± 7.0 years) as compared to controls (2.8 ± 5.0 years, P-0.044).

Table 3.

Risk Factors for Nonalcoholic Fatty Liver Disease (NAFLD) in Middle-aged Subjects and Elderly Subjects.

Risk factors	Middle age (40 yrs–59 yrs)			Elderly (≥60 yrs)
	Cases (n-76)	Controls (n-41)	P value	Cases (n-42)	Controls (n-18)	P value
Dyslipidemia∗ {n (%)}	16 (21%)	06 (15%)	0.397	09 (24%)	05 (28%)	0.745
Family history of dyslipidemia {n (%)}	06 (08%)	05 (12%)	0.447	04 (9.5%)	03 (17%)	0.430
Diabetes mellitus {n (%)}	38 (50%)	13 (32%)	0.057	23 (55%)	03 (17%)	0.006
Family history of DM {n (%)}	18 (25%)	06 (15%)	0.192	17 (40%)	02 (11%)	0.025
Obesity {n (%)}	38 (51%)	16 (39%)	0.204	25 (57%)	07 (39%)	0.195
Family history of obesity {n (%)}	17 (22%)	02 (05%)	0.014	08 (19%)	00 (0%)	0.047
Hypertension {n (%)}	16 (21%)	09 (22%)	0.910	16 (38%)	06 (33%)	0.726
Family history of hypertension {n (%)}	07 (9%)	02 (05%)	0.401	06 (14%)	02 (11%)	0.740
Metabolic syndrome {n (%)}	45 (59%)	18 (44%)	0.113	32 (76%)	08 (44%)	0.017
Total cholesterol (mg/dl, mean ± SD)	122 ± 45	110 ± 42	0.192	126 ± 40	118 ± 54	0.522
Triglycerides (mg/dl, mean ± SD)	90 ± 41	77 ± 43	0.120	89 ± 33	86 ± 39	0.787
High-density lipoprotein (mg/dl, mean ± SD)	37 ± 15	33 ± 17	0.181	34 ± 15	36 ± 19	0.683
Low-density lipoprotein (mg/dl, mean ± SD)	71 ± 35	68 ± 30	0.668	76 ± 31	71 ± 42	0.592
Glucose-fasting (mean ± SD) mg/dl	116 ± 31	118.6 ± 54	0.763	118 ± 37	100 ± 21	0.058
Glucose-postprandial (mean ± SD) mg/dl	182 ± 83	178.4 ± 82.08	0.810	182 ± 65	153 ± 47	0.087
Duration of DM in years{median (IQR)}	6 (0)	5 (0)	0.376	10 (1)	7 (0)	0.151
Weight {kg, median (range)}	67.6 (40.3–100.8)	64 (45–83.5)	0.109	67.15 (42–95)	62.5 (46–90.1)	0.390
Waist-hip ratio {median (range)}	1.02 (0.89–2.0)	1.02 (0.91–1.11)	0.532	1.02 (0.91–1.17)	1.03 (0.97–1.03)	0.451
BMI∗ {kg/m2, median (range)}	25.1 (16.6–35.7)	23.6 (18.8–32.9)	0.085	25.3 (19–35.3)	23.65 (17.3–31.1)	0.171
Lifetime weight as obese (years, mean ± SD)	5.5 ± 5.7	3.3 ± 5.2	0.045	6.6 ± 7.0	2.8 ± 5.0	0.044
Premorbid weight (kg, mean ± SD)	69.1 ± 13.5	63.3 ± 9.2	0.015	70.3 ± 13.1	64.0 ± 12.1	0.087
≥3 risk factors {n (%)}	18 (25%)	09 (22%)	0.411	15 (36%)	05 (28%)	0.102

The number of risk factors for NAFLD among cases and controls is depicted in Figure 1. In middle age group, 77% of cases and 68% of controls (P-0.148) had one or more risk factors for NAFLD, whereas ≥3 risk factors were seen in 25% cases and 22% controls (P-0.411). Among elderly, 90% of cases and 84% of controls (P-0.813) had one or more risk factors for NAFLD and ≥3 risk factors were seen in 38% of cases and 28% of controls (P-0.102). This has been shown in Figure 3a and 3b.

Figure 3.

(a) Number of risk factors for NAFLD in elderly patients. (b) Number of risk factors for NAFLD in middle-aged patients.

Multivariate analysis was performed to evaluate further the association of risk factors with cryptogenic cirrhosis. Independent variables included in the model were diabetes mellitus, metabolic syndrome, family history of diabetes mellitus, family history of obesity, premorbid weight and lifetime weight as obese (risk factors significantly associated in univariate analysis) to determine if these had independent significant effects on the dependent variable (cryptogenic cirrhosis). None of the risk factors showed a significant association with cryptogenic cirrhosis.

Liver Biopsy

Liver biopsy was available in 12 (10%) cases (4 were explants) and in 2 (3%) controls (both explants). In majority of cases (11/12) there was evidence of steatosis in liver biopsy. Nine of the 12 cases (75%) were having 3 or more risk factors for NAFLD. One case with steatosis on liver biopsy had only one risk factor, whereas two of the cases were not having any risk factors and one had steatosis on liver biopsy.

Occult Hepatitis B Infection

All patients with cryptogenic cirrhosis (cases) underwent total anti hepatitis B core antibody testing to look for occult HBV infection. This was positive in 22 (19%) cases. HBV DNA results, although advised in all anti-hepatitis B core antibody positive patients, were done only by 5 and all were negative.

Discussion

In our study, we noted that risk factors for NAFLD were more prevalent in cases of cryptogenic cirrhosis than in controls with HBV- or HCV-related cirrhosis. The duration of diabetes mellitus and obesity were longer in cases compared to controls. Anti-hepatitis B core antibody was positive in 22 (19%) cases; HBV DNA was tested in 5 patients and was negative in all.

In our study, the incidence of ascites (53% vs 44%), GI bleed (15% vs 8.5%), spontaneous bacterial peritonitis (9% vs 12%), acute kidney injury (6% vs 5%) and hepatic encephalopathy (36% vs 26%) were similar in the two groups. As majority of the patients had ascites, calculation of BMI, waist-hip ratio and waist circumference may not be accurate and this is one of the limiting factor of our study. Rinaldi et al also reported comparable percentages [ascites (54.9% vs 51.8%), encephalopathy (25.5% vs 21.81%), oesophageal variceal bleed (6.9% vs 6.3%) and HCC (16.7% vs 17.2%)] in both cases and controls.¹⁴ Caldwell et al reported ascites in 30% of patients, portal hypertensive bleed in 20% and hepatic encephalopathy in 6%.¹⁵ Marmur et al however reported variceal bleeding in 33%, ascites in 89%, spontaneous bacterial peritonitis in 10% and hepatic encephalopathy in 23%, probably because of the differences in the study population.¹⁶ Five percent of cases had hepatocellular carcinoma in our study. Caldwell et al reported a prevalence of 1.4%,¹⁵ whereas Kojima et al reported a prevalence of 39% (9 out of 23) and Ratziu et al 27% (8 out of 27), in the cryptogenic group.^17,18

In Indian patients with cryptogenic cirrhosis, previous studies have shown that NAFLD, autoimmune hepatitis, occult HBV infection and noncirrhotic fibrosis are the common causes.¹⁹ In a previous study from our institute, 35% of the patients had cryptogenic cirrhosis and it was the commonest aetiology for portal hypertension in adults. Among 203 patients with cryptogenic cirrhosis 39 patients underwent liver biopsy and on biopsy 16 patients had idiopathic non cirrhotic intrahepatic portal hypertension (NCIPH), 8 patients had autoimmune hepatitis, 5 patients each had NAFLD and cryptogenic cirrhosis. Hepatic amyloidosis and Wilson's disease were noted in 1 each. In patients without liver biopsy, 77% had risk factors for NAFLD and criteria for probable autoimmune hepatitis were satisfied in 5 patients. Although it was a prospective study, the authors looked retrospectively for risk factors of NAFLD in patients with cryptogenic cirrhosis who did not undergo liver biopsy. One of the important limitations of the study according to the authors was that the risk factors of NAFLD were not assessed in all patients.⁵ Even in our study one or more risk factors for NAFLD was noted in 79% of cases and ≥3 risk factors were seen in 28% cases.

Diabetes mellitus was more prevalent among cases than controls (52% vs 27%, P-0.002). Although not significant, cases had a longer duration of diabetes than controls (median/IQR – 7 ± 0.5 years v/s 5 ± 0 years, P – 0.113). Caldwell et al found the prevalence of obesity (47% vs 3%) and type II diabetes mellitus (53% vs 25%) higher in cases compared to controls. However, the prevalence of obesity (64% v/s 47%) and type II diabetes mellitus (42% v/s 53%) was similar in patients with NASH and cryptogenic cirrhosis.¹⁵ Poonawala et al also reported a higher prevalence of obesity (55% vs 24%) and diabetes mellitus (47 vs 22%) in patients with cryptogenic cases compared to controls, as did Caldwell et al.^4,15

Obesity was more common among cases than controls (53% vs 39%, P-0.081). Duration of obesity (lifetime weight as obese) was longer in cases (5.9 ± 6.2 years) as compared to controls (3.2 ± 5.1 years, P-0.002). Tellez-Avila et al also showed a higher prevalence of obesity (16.4% v/s 8.2%) and diabetes (40% vs 22.4%) in Mexican patients with cryptogenic cirrhosis compared to controls.⁶ Kojima et al reported prevalence of obesity in 54.2%, diabetes mellitus in 54.2% and hypertriglyceridemia in 20.8% of Japanese patients with cryptogenic cirrhosis which was higher compared to controls.¹⁷ In another study from our institution in patients with non B non C hepatocellular carcinoma, 85% of patients had at least one risk factor for NAFLD, 15% of patients were overweight, 51% of patients were obese and their lifetime body weight as obese was 8.9 years. Among patients who were obese, 75% of patients had at least one risk factor for NAFLD.²⁰

Metabolic syndrome was noted in 65% of cases, as compared to 44% of controls (P-0.001) in our study and it was similar to the study by Rajan et al (70.6% vs 26.2%; P < 0.001).¹ Tellez-Avila et al also reported a higher prevalence of metabolic syndrome in Mexican patients with cryptogenic cirrhosis.⁶

Family history of diabetes mellitus (30% vs 13%, P-0.016) and obesity (21% vs 3.5%, P-0.002) was more common among cases than controls. Bhadoria et al had also shown a higher incidence of metabolic traits in first degree relatives of patients with NASH or cryptogenic cirrhosis. About 68.8% patients had family history of atleast one of the metabolic traits, higher prevalence of family history of diabetes (52.5%), hypertension (46.2%), coronary artery disease (21.2%) and obesity (53.3%).²¹ Another study by the same author, on alcoholic liver disease, revealed that 63.5% had family history of one or more metabolic traits.²² In both studies, presence of family history of metabolic traits was associated with earlier age of cirrhosis, higher MELD score and early decompensation.^21,22

In our study, the risk factors for NAFLD like DM, metabolic syndrome, family history of obesity and family history of DM were more common in cases compared to controls. The duration of DM and lifetime weight as obese was also more in cases. In a study by Duseja et al among 65 patients with cryptogenic cirrhosis and 50 virus-related cirrhosis patients, patients with cryptogenic cirrhosis had a higher mean BMI, more abnormal waist, more type 2 DM and a lower serum HDL. Similarly among 39 patients with cryptogenic HCC and 39 patients with virus related HCC, patients with cryptogenic HCC had higher BMI and higher prevalence of type 2 DM.²³

In another study from India, cryptogenic cirrhosis occurred more commonly in patients with DM and patients with cryptogenic cirrhosis had higher incidence of DM compared to patients with non-cryptogenic cirrhosis.²⁴

In our study elderly patients with cryptogenic cirrhosis had more risk factors for NAFLD than middle-aged patients. In elderly age group, DM (55% v/s 17%, P-0.006), family history of DM (40% v/s 11%, P-0.025), metabolic syndrome (76% v/s 44%, P-0.017) and family history of obesity (19% v/s 0, P-0.047) were more common in cases as compared to controls, whereas in the middle age group,family history of obesity was the only significant factor (22% v/s 5%, P-0.025). In general population, elderly patients have more prevalence of DM, hypertension, obesity and metabolic syndrome, but there are limited data available regarding risk factors for NAFLD in elderly cryptogenic cirrhosis patients and we need more studies in this population.25, 26, 27

Liver biopsy was available in 12 (10%) cases (4 were explants) and 2 (3%) controls (both explants). Most biopsies (11 out of 12) showed evidence of steatosis, 75% of the cases (9/12) had ≥ 3 risk factors for NAFLD. Poonawala et al noted steatosis on liver histology in 10% of cases and 7% of controls.⁴ Bansal et al noted 5% of the patients (9/180) had NAFLD-related cirrhosis with 3 having HCC and 8% (15/180) had cryptogenic cirrhosis, with 4 having HCC.²⁸ In the study by Kojima et al, liver biopsy was done in 12 of 24 patients with cryptogenic cirrhosis; 10 of these showed one or more features suggestive of nonalcoholic steatohepatitis.¹⁷ Contos et al studied liver biopsies in patients who had undergone liver transplantation for cryptogenic cirrhosis and they reported significant rate of reappearance of steatosis in allografts which were almost 100% in 5 years. This was much higher compared to allografts of alcoholics and those with primary biliary cirrhosis/primary sclerosing cholangitis which was only 25% at 5 years.²⁹ In an explant study from South India, 89.6% of the patients showed concordance with pretransplant diagnosis, and it was 100% among patients with alcohol-related liver disease, hepatitis B/hepatitis C related liver disease, autoimmune liver disease, biliary cirrhosis and Budd-Chiari syndrome. In patients with cryptogenic cirrhosis (n – 37), discordance was seen in 23 patients, 5 patients had haemochromatosis, 7 patients each had autoimmune hepatitis and NASH, 2 patients had noncirrhotic fibrosis, Wilson’s disease and congenital hepatic fibrosis was noted in 1 patient each. Among patients who were diagnosed to have NASH (7/37) on explant did not have any of the components of metabolic syndrome and there median BMI was 23 kg/m² and this is in contrast to findings in our study.¹⁹

All cryptogenic cirrhosis cases underwent anti-hepatitis B core antibody assay to look for occult HBV infection. Anti-HBc antibody was positive in 22 (19%) cases. HBV DNA was tested in 5 of anti-HBc antibody positive patients and was negative. Radhakrishnan et al noted serum HBV DNA positivity in only 3.3% patients with cryptogenic cirrhosis.³⁰ Chan et al in their study noted that 24 out of 28 patients with cryptogenic cirrhosis had anti-HBc antibody positivity, while in these only 9 had HBV DNA positivity (32%) denoting occult HBV infection.³¹

One of the major limitation of this study is that the study data is seven years old. However, in our opinion these data deserve to be highlighted, as even today, there is a paucity of studies from India providing a detailed analysis of risk factors of NAFLD in patients with cryptogenic cirrhosis. It is also important to note that there is no data currently available in elderly cryptogenic cirrhosis patients, which is a subgroup, where the risk factors are considered much more prevalent than in the younger age groups. Other limitation of this study includes possible referral bias, since this was a single centre study done in a tertiary care hospital. The study also had a limited sample size. Not all anti-HBc positive patients underwent serum HBV DNA estimation and only a limited number underwent liver biopsy. Recall bias was also a potential limitation in this study especially for the variables such as duration of diabetes, lifetime body weight and premorbid body weight. Also, in patients with ascites, body mass index, waist circumference and waist hip ratio measurements may not have been accurate.

Conclusion

This study found a significantly higher prevalence of risk factors for NAFLD in middle aged and elderly patients with cryptogenic cirrhosis when compared to patients with hepatitis B or C related cirrhosis. The duration of lifetime weight as obese was also significantly longer in patients with cryptogenic cirrhosis. Elder patients with cryptogenic cirrhosis had more risk factors for NAFLD than middle-age group subjects.

Credit authorship contibution statement

B.K.C. contributed to study concept and design, collection of data, interpretation of data, critical revision of the manuscript, and approved the final submission. A.G. contributed to interpretation of data, critical revision of the manuscript, and approved the final submission. R.J. contributed to interpretation of data, critical revision of the manuscript, and approved the final submission. D.D. contributed to collection of data, interpretation of data, critical revision of the manuscript, and approved the final submission. U.Z. contributed to interpretation of data, critical revision of the manuscript, and approved the final submission. J.R. contributed to study concept and design, contributed to interpretation of data, critical revision of the manuscript, and approved the final submission. C.E.E. contributed to study concept and design, contributed to interpretation of data, critical revision of the manuscript, and approved the final submission.

Conflicts of interest

The authors have none to declare.

Acknowledgement

The authors are grateful for funds received from the Fluid Research Fund at Christian Medical College, Vellore, India for this study.

Funding

None.