Abstract
Background
Previous research has demonstrated the clinical significance of alanine aminotransferase (ALT) flares; however, studies have mostly been in Asian countries or populations, and it remains unclear whether ALT flares during treatment are associated with hepatitis B surface antigen (HBsAg) loss and long-term adverse clinical outcomes.
Objectives
To evaluate the association between ALT flares and virologic outcomes (HBsAg and hepatitis B e antigen [HBeAg] loss) as well as adverse clinical outcomes among patients with chronic hepatitis B in the United States, according to treatment status.
Design
Retrospective study using the Optum de-identified electronic health record dataset (2012–2019).
Methods
Marginal structural models estimated the associations between ALT flares and outcomes, accounting for time-varying confounding; adjusted odds ratios and 95% confidence intervals were reported. A Cox proportional hazards regression model was used to assess risk factors for flares.
Results
14,328 patients were included in the untreated cohort; of these, 2298 (16.0%) initiated and 1541 (10.7%) subsequently discontinued treatment. At least one ALT flare was experienced by 364 patients (2.5%) in the untreated cohort, 84 (3.7%) in the treatment initiation cohort, and 22 (1.4%) in the discontinuation cohort. Risk factors for ALT flares in the untreated group included male sex, history of flares, metabolic syndrome, liver fibrosis, compensated cirrhosis (CC), and hepatic decompensation. Risk factors after treatment initiation included younger age, White race, history of flares, and evidence of liver damage (liver fibrosis, CC, or hepatic decompensation). Flares in the untreated group were associated with spontaneous HBsAg loss and with an increased risk of hepatic decompensation, hospitalization, and death. Flares after treatment initiation were associated with HBsAg and HBeAg loss but not with adverse clinical outcomes investigated.
Conclusion
ALT flares in untreated patients were associated with virologic and adverse clinical outcomes; no association with adverse clinical outcomes was observed in patients who initiated treatment.
Keywords: ALT flare, chronic HBV, clinical outcomes, virologic outcomes
Introduction
Despite progress achieved in disease prevention and treatment in recent decades, chronic hepatitis B virus (HBV) infection continues to represent a major public health concern both in the United States (US) and globally.1,2 In the US in 2022, an estimated 1.65 million people were living with chronic HBV infection. 1 Chronic HBV infection can progress to serious, life-threatening complications, including cirrhosis, hepatic decompensation, and hepatocellular carcinoma.3,4 In 2022, an estimated 1.1 million HBV-related deaths were reported globally 5 ; 1769 HBV-related deaths were reported in the US in 2023. 6 In 2019, 40% of the global deaths from liver cancer and 23% of deaths from cirrhosis were HBV-related. 7
Increases (or flares) in serum alanine aminotransferase (ALT) levels are commonly observed in patients with chronic HBV infection. 8 Various definitions of ALT flares exist, although an abrupt increase in serum ALT greater than 5× the upper limit of normal (ULN) is the minimum requirement among the most common definitions.9–11 ALT flares can occur during both hepatitis B envelope antigen (HBeAg)-positive and -negative immune-active phases of HBV infection. 8 Spontaneous elevations of serum ALT levels are often associated with the onset of the host cytolytic immune response to HBV infection and ensuing lysis of infected hepatocytes 8 ; serum ALT flares, therefore, reflect intrahepatic necro-inflammatory activity and provide a non-invasive measure of disease activity.8,12 The severity and frequency of ALT flares have been shown to correlate with the risk of cirrhosis and hepatocellular carcinoma (HCC) development. 8
Current treatment options for patients with chronic HBV infection are long-term nucleos(t)ide analog (NA) therapy or interferons (IFNs),12–14 and ALT flares can occur in the context of anti-HBV treatment. These flares can follow a benign course or lead to severe hepatitis/hepatic decompensation. 8 Flares during treatment have been beneficially associated with declines in serum HBV DNA and hepatitis B surface antigen (HBsAg), leading to an accelerated serologic response. 9 However, HBsAg loss with current treatments is rare (1–4% with NA, 2–8% with pegylated IFN [PEG-IFN]).2,15–17 Flares are also common after withdrawal of NA treatment, especially within the first year, and may result in hepatic decompensation and death. 18
Previous research has demonstrated the clinical significance of ALT flares; however, studies have mostly been conducted in Asian countries or populations,19–21 and it is still unclear whether ALT flares during treatment are associated with HBsAg loss and long-term clinical outcomes such as cirrhosis, hepatic decompensation, and HCC. It is also important to supplement existing literature with data from ethnically diverse populations to establish generalizable evidence and to inform clinical decisions for patients with chronic HBV infection. The main objective of this study was to assess the association between the onset of ALT flares and both virologic (i.e., HBsAg and HBeAg loss) and adverse clinical outcomes (i.e., hepatic decompensation, all-cause hospitalization, and all-cause mortality) among a diverse US population infected with chronic hepatitis B in relation to their treatment status (untreated, after treatment initiation, and after treatment discontinuation). We also describe the risk factors associated with, and dynamics of, ALT flares.
Methods
The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (https://www.strobe-statement.org/). 22
This retrospective cohort study used data from the Optum electronic health record (EHR) database. This database aggregates data from approximately 140,000 providers at more than 700 hospitals and 7000 clinics in the US. Approximately 97 million patients are included from all US Census regions, contributing to its generalizability to the US population receiving routine clinical care. A nested cohort study design was applied, as described below, whereby an initial untreated chronic HBV cohort was followed up until treatment initiation, giving rise to the treatment initiation cohort, with this cohort followed up until discontinuation, giving rise to the treatment discontinuation cohort. This ensured, notwithstanding left censoring, the selection of incident initiators and discontinuers to appropriately capture first ALT flares.
Study population
The study population consisted of adult patients (≥18 years old) diagnosed with chronic HBV infection between January 1, 2012, and December 31, 2018. Chronic HBV infection was defined as any combination of “positive HBsAg result” or “positive HBV DNA” (including genotype testing) or “positive HBeAg” on ≥2 occasions ≥6 months apart, with the latter considered the point of chronicity, or presence of at least one International Classification of Diseases (ICD)-9/ICD-10 diagnosis code for chronic HBV infection.
Patients were excluded if they had missing data on age and/or sex, ongoing ALT flare at study entry (defined as ongoing if the period between ALT flare start and end dates overlapped with the study entry index date), any history of a negative HBsAg test result prior to or on the study entry index date, and any history of treatment with IFNs or NAs prior to or on the study entry index date. A full list of the inclusion and exclusion criteria, which were applied at the respective index date for each study cohort (untreated, treatment initiation, and treatment discontinuation), is provided in Supplemental Table 1.
A cohort of patients who initiated anti-HBV treatment was identified from the untreated cohort above. The treatment initiation index date was defined as the date of the first-ever prescription of IFN, PEG-IFN, or NAs (entecavir [ETV], tenofovir disoproxil fumarate [TDF], or tenofovir alafenamide fumarate [TAF]) after the patient’s study entry index date. Among those who initiated treatment, another cohort of patients who discontinued treatment was analyzed (Figure 1). A treatment discontinuation event was defined as a change in status from treated to untreated (i.e., not receiving any anti-HBV treatment). If a gap of >90 days was observed without a new antiviral prescription or other anti-HBV treatment, the date of treatment discontinuation was the end date of the last continuous prescription. This gap was part of the treatment discontinuation algorithm (rather than an inclusion/exclusion criterion).
Figure 1.
Study design. ALT, alanine aminotransferase; HBV, hepatitis B virus; IFN, interferon; NA, nucleos(t)ide analog
Patients in each cohort were followed up from their respective index dates (untreated, study entry; treatment initiation; treatment discontinuation) to the last active month, first ALT flare after a treatment transition (initiation for the untreated cohort, discontinuation for the treatment initiation cohort, or re-initiation for the treatment discontinuation cohort), or death. ALT flare was defined as ALT >5 × ULN and ALT >2 × the baseline or nadir ALT value; a fixed value of 40 U/L was used for ULN. The lowest ALT value (i.e., baseline or nadir after index) was taken as the reference ALT value to identify the first ALT flare. The lowest ALT value (i.e., baseline corresponding to the ALT measurement at the end of the previous flare or nadir after the end of the previous flare) was taken as the reference value to identify subsequent ALT flares. ALT >2 × baseline or nadir definition was used to rule out potential transient fluctuations in ALT levels. The end date of an ALT flare was the date of the first ALT value after the start of the ALT flare that did not meet both parts of the ALT definition (i.e., ≤5 × ULN and ≤2 × baseline or nadir), or the date 12 weeks following the last available ALT measurement, whichever came first.
Outcome definitions
Virologic outcomes
HBsAg loss was defined as having two consecutive negative HBsAg tests ≥6 months apart, with the latter date taken as the date of loss; only patients with a positive HBsAg test and no prior evidence of HBsAg negativity at study entry index date, treatment initiation, or discontinuation index date were considered at risk of HBsAg loss for each cohort. HBeAg loss was defined as one negative HBeAg test following a positive HBeAg test; only patients with a positive HBeAg test at the study entry index date, treatment initiation, or discontinuation index date were considered at risk of HBeAg loss for each cohort.
Clinical outcomes
Hepatic decompensation was defined by the presence of diagnosis codes for any of ascites, variceal bleeding, hepatic encephalopathy, hepatic coma, or spontaneous bacterial peritonitis. All-cause hospitalization was defined as an overnight hospital stay due to any cause. All-cause mortality was defined as death due to any cause.
Statistical analysis
Baseline characteristics
Demographic, clinical, and treatment characteristics were summarized by descriptive statistics at different time points in patients experiencing a flare (henceforth “flare patients”) or not experiencing a flare (henceforth “no-flare patients”), i.e., at study entry index, treatment initiation index, treatment discontinuation index, and at the onset of the first ALT flare.
Any evidence of advanced liver disease was defined as ≥1 aspartate aminotransferase (AST) to platelet ratio index (APRI) score >2.0 or Fibrosis-4 (FIB-4) score >3.25 in the 12-month period prior to or on the respective index date (study entry, treatment initiation, and treatment discontinuation). Metabolic syndrome indicated the presence of overweight/obesity, type 2 diabetes mellitus, hypertension, and hyperlipidemia.
Occurrence of and risk factors for ALT flares
Incidence rates (IRs) of ALT flares in each cohort per 1000 person-years (PY) were calculated as the total number of ALT flares divided by the total PY follow-up time. Time from study entry, treatment initiation, and treatment discontinuation index to the first ALT flare experienced by patients in each cohort was described using Kaplan–Meier curves.
Factors predictive of experiencing ALT flares were evaluated. The associations between selected patient demographic, clinical, biochemical, and treatment characteristics and onset of the first ALT flare among untreated patients and among those who initiated treatment were investigated using time-dependent Cox proportional hazards regression models. Covariates were included in the models as time-fixed or time-varying variables, and the hazard ratio (HR) and associated 95% confidence interval (CI) for each were calculated.
Characteristics of ALT flares and their relationship with other laboratory parameters
Characteristics of flares (the number of patients with ≥1 ALT flare, the IR of ALT flares, time to start of first ALT flare, peak value of first ALT flare, and duration of the first ALT flare) in the overall cohort and in treatment initiation/discontinuation cohorts were described. Continuous variables were described by the mean, standard deviation (SD), median, and 25th and 75th percentiles. Categorical variables were described by the number and percentage of patients in each category among those with available data. The number of patients with missing data for each variable was reported.
Among all three cohorts, the locally estimated scatterplot smoothing method was used to plot the curves of ALT levels over time for patients experiencing ALT flares with the curves of other biochemical and virologic measurements (i.e., AST, alkaline phosphatase, serum bilirubin, platelet counts, HBV DNA) during the same time periods.
Association between ALT flares and long-term clinical outcomes
Marginal structural Cox proportional hazards models 23 were used to estimate the associations between ALT flares in the untreated and treatment initiation cohorts and the outcomes of interest. The planned marginal structural model analysis for ALT flares among the discontinuation cohort was not conducted due to the small sample size. For each virologic/clinical outcome-specific analysis, patients with a history of that outcome prior to or on the respective index date for the untreated cohort and treatment initiation cohort were excluded. Both time-fixed and time-varying covariates (evaluated and updated in each 30-day interval of follow-up) were accounted for in the regression models for inverse probability of treatment weighting (IPTW). Competing risk (i.e., all-cause mortality for non-death outcomes) and censoring due to first ALT flare after treatment initiation/discontinuation/re-initiation for all outcomes were accounted for through separate inverse probability of censoring weighting (IPCW), adjusting for both time-fixed and time-varying covariates. A weighted pooled logistic regression model was fitted to approximate a Cox proportional hazards model and estimate HRs for the association between ALT flares and outcomes of interest, while incorporating IPTW and IPCW. Odds ratios (ORs; approximating HRs) with 95% CIs were reported, and adjusted cumulative survival curves for patients experiencing ALT flares vs no flares were plotted against time for each outcome.
Results
Patient attrition
In total, 14,328 patients were included in the initial untreated cohort (Figure 2). Among these, 2298 patients (16.0%) initiated treatment; within this group, 1708 patients (74.3%) discontinued treatment during follow-up. Within this initial treatment discontinuation cohort, 106 patients were flagged as discontinued due to loss of follow-up or death and a further 167 patients did not meet the eligibility criteria for the discontinuation cohort (Supplemental Table 4, with patient attrition for each cohort included in Supplemental Tables 2–4). These patients were excluded, resulting in 1541 patients for analysis in the discontinuation cohort.
Figure 2.
Patient attrition. Each inclusion criterion step is sequential (i.e., proportion of patients is out of the number of patients in the previous step). aThe proportion is out of the number of patients in the treatment initiation cohort (n = 2298). Among 1602 patients who discontinued treatment due to a ≥90-day gap, 1057 (66.0%) reinitiated treatment during follow-up. bCounts and proportions of patients meeting each exclusion criterion are out of the number of patients who met all inclusion criteria (n = 1708) and are not mutually exclusive. In total, 167 distinct patients were excluded. cAny history was defined as the full period prior to the index date. dAny evidence was defined as 1 year prior to the index date. 3TC, lamivudine; ADV, adefovir dipivoxil; ALT, alanine aminotransferase; ETV, entecavir; HAV, hepatitis A virus; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; HCV, hepatitis C virus; HDV, hepatitis D virus; HIV, human immunodeficiency virus; IFN, interferon; LdT, telbivudine; NA, nucleos(t)ide analog; PEG-IFN, pegylated IFN; TAF, tenofovir alafenamide fumarate; TDF, tenofovir disoproxil fumarate.
Occurrence of ALT flares
At least one ALT flare was experienced by 364 (2.5%; 95% CI 2.3–2.8%) of the 14,328 patients in the untreated cohort, 84 (3.7%; 95% CI 2.9–4.5%) of 2298 patients in the treatment initiation cohort, and 22 (1.4%; 95% CI 0.9–2.2%) of 1541 patients in the treatment discontinuation cohort (Supplemental Table 5). Most patients who experienced an ALT flare had only one recorded during the follow-up period.
The unadjusted IR of ALT flare was 7.9 per 1000 PY in the untreated cohort, 29.6 per 1000 PY in the treatment initiation cohort, and 11.8 per 1000 PY in the treatment discontinuation cohort (Table 1).
Table 1.
Characteristics of ALT flares in the untreated cohort, treatment initiation cohort, and treatment discontinuation cohort.
| | Untreated cohort (n = 14,328) | Treatment initiation cohort (n = 2298) | Treatment discontinuation cohort (n = 1541) |
|---|---|---|---|
| Total PY | 50,748.9 | 3003.6 | 2115.5 |
| Number of ALT flares | 402 | 89 | 25 |
| IR per 1000 PY (95% CI) | 7.9 (7.1–8.8) | 29.6 (23.8–36.9) | 11.8 (7.5–18.5) |
| Peak value of first flare, U/L (continuous) | |||
| N | 364 | 84 | 22 |
| Mean (SD) | 652.5 (773.8) | 565.6 (476.5) | 516.8 (850.1) |
| 95% CI for the mean | 573.0–732.0 | 463.7–667.5 | 161.6–872.0 |
| Median (Q1–Q3) | 375 (259–685) | 391 (269–671) | 272 (238–389) |
| Peak value of first flare (categorical), n (%) [95% CI] | |||
| >5 × to ≤10 × ULN | 201 (55.2) [50.0–60.4] | 43 (51.2) [40.0–62.3] | 17 (77.3) [54.6–92.2] |
| >10 × to ≤20 × ULN | 85 (23.4) [19.1–28.0] | 26 (31.0) [21.3–42.0] | 3 (13.6) [2.9–34.9] |
| >20 × ULN | 78 (21.4) [17.3–26.0] | 15 (17.9) [10.4–27.7] | 2 (9.1) [1.1–29.2] |
| Time to first flare, days | |||
| N | 364 | 84 | 22 |
| Mean (SD) | 635.1 (732.9) | 200.4 (367.8) | 249.0 (238.6) |
| Median (Q1–Q3) | NR | NR | NR |
| Duration of first flare, days | |||
| N | 364 | 84 | 22 |
| Mean (SD) | 108.5 (91.6) | 94.0 (58.4) | 94.7 (37.6) |
| 95% CI for the mean | 99.1–118.0 | 81.5–106.5 | 79.0–110.4 |
| Median (Q1–Q3) | 86 (85–119) | 85 (60–111) | 85 (85–108) |
ALT, alanine aminotransferase; CI, confidence interval; IR, incidence rate; NR, not reported; PY, person-years; Q, quartile; SD, standard deviation; ULN, upper limit of normal.
Patient characteristics
Untreated cohort
The mean age at study entry index was 45.9 years in both flare and no-flare patients (Table 2). The proportion of males was higher among flare than no-flare patients (69.0% [95% CI 63.9–73.7%] vs 52.4% [95% CI 51.6–53.2%]), as was the proportion of non-Hispanic Caucasian (24.2% [95% CI 19.9–28.9%] vs 15.9% [95% CI 15.3–16.5%]) and non-Hispanic Black/African American (20.3% [95% CI 16.3–24.8%] vs 13.6% [95% CI 13.1–14.2%]).
Table 2.
Patient characteristics at respective index dates.
| Variable | Untreated cohort (n = 14,328) | Treatment initiation cohort (n = 2298) | Treatment discontinuation cohort (n = 1541) | No treatment initiation during follow-up (n = 12,030) | |||
|---|---|---|---|---|---|---|---|
| ALT flares | No flares | ALT flares | No flares | ALT flares | No flares | ||
| N; % (95% CI) a | N; % (95% CI) a | N; % (95% CI) a | N; % (95% CI) a | N; % (95% CI) a | N; % (95% CI) a | N; % (95% CI) a | |
| Demographics | |||||||
| Age (continuous) | |||||||
| N | 364; 100 | 13,964; 100 | 84; 100 | 2214; 100 | 22; 100 | 1519; 100 | 12,030; 100 |
| Mean (SD), years | 45.9 (12.6) | 45.9 (13.2) | 46.3 (13.9) | 48.3 (13.0) | 49.7 (13.0) | 48.8 (13.1) | 45.7 (13.2) |
| Median (Q1–Q3) | 45.5 (37.0–55.0) | 45.0 (36.0–55.0) | 47.0 (37.5–54.5) | 48.0 (39.0–57.0) | 47.0 (41.0–63.0) | 49.0 (39.0–58.0) | 44.0 (36.0–55.0) |
| Age (categorical) | |||||||
| 18–<30 years | 41; 11.3 (8.2–15.0) | 1372; 9.8 (9.3–10.3) | 11; 13.1 (6.7–22.2) | 133; 6.0 (5.1–7.1) | 2; 9.1 (1.1–29.2) | 88; 5.8 (4.7–7.1) | 1231; 10.2 (9.7–10.8) |
| 30–<40 years | 75; 20.6 (16.6–25.1) | 3534; 25.3 (24.6–26.0) | 17; 20.2 (12.3–30.4) | 479; 21.6 (19.9–23.4) | 2; 9.1 (1.1–29.2) | 315; 20.7 (18.7–22.9) | 3058; 25.4 (24.6–26.2) |
| 40–<50 years | 110; 30.2 (25.5–35.2) | 3789; 27.1 (26.4–27.9) | 21; 25.0 (16.2–35.6) | 598; 27.0 (25.2–28.9) | 9; 40.9 (20.7–63.6) | 393; 25.9 (23.7–28.2) | 3286; 27.3 (26.5–28.1) |
| 50–<60 years | 80; 22.0 (17.8–26.6) | 2914; 20.9 (20.2–21.6) | 19; 22.6 (14.2–33.0) | 546; 24.7 (22.9–26.5) | 3; 13.6 (2.9–34.9) | 405; 26.7 (24.5–29.0) | 2461; 20.5 (19.7–21.2) |
| 60–<70 years | 48; 13.2 (9.9–17.1) | 1718; 12.3 (11.8–12.9) | 12; 14.3 (7.6–23.6) | 334; 15.1 (13.6–16.6) | 4; 18.2 (5.2–40.3) | 226; 14.9 (13.1–16.8) | 1453; 12.1 (11.5–12.7) |
| ≥70 years | 10; 2.7 (1.3–5.0) | 637; 4.6 (4.2–4.9) | 4; 4.8 (1.3–11.7) | 124; 5.6 (4.7–6.6) | 2; 9.1 (1.1–29.2) | 92; 6.1 (4.9–7.4) | 541; 4.5 (4.1–4.9) |
| Sex | |||||||
| Female | 113; 31.0 (26.3–36.1) | 6648; 47.6 (46.8–48.4) | 22; 26.2 (17.2–36.9) | 917; 41.4 (39.4–43.5) | 5; 22.7 (7.8–45.4) | 635; 41.8 (39.3–44.3) | 5822; 48.4 (47.5–49.3) |
| Male | 251; 69.0 (63.9–73.7) | 7316; 52.4 (51.6–53.2) | 62; 73.8 (63.1–82.8) | 1297; 58.6 (56.5–60.6) | 17; 77.3 (54.6–92.2) | 884; 58.2 (55.7–60.7) | 6208; 51.6 (50.7–52.5) |
| Race/ethnicity | |||||||
| Non-Hispanic Caucasian | 88; 24.2 (19.9–28.9) | 2219; 15.9 (15.3–16.5) | 27; 32.1 (22.4–43.2) | 327; 14.8 (13.3–16.3) | 2; 9.1 (1.1–29.2) | 211; 13.9 (12.2–15.7) | 1953; 16.2 (15.6–16.9) |
| Hispanic Caucasian | 4; 1.1 (0.3–2.8) | 148; 1.1 (0.9–1.2) | 3; 3.6 (0.7–10.1) | 16; 0.7 (0.4–1.2) | 1; 4.5 (0.1–22.8) | 12; 0.8 (0.4–1.4) | 133; 1.1 (0.9–1.3) |
| Non-Hispanic Asian | 130; 35.7 (30.8–40.9) | 5565; 39.9 (39.0–40.7) | 29; 34.5 (24.5–45.7) | 1032; 46.6 (44.5–48.7) | 10; 45.5 (24.4–67.8) | 691; 45.5 (43.0–48.0) | 4634; 38.5 (37.6–39.4) |
| Hispanic Asian | 1; 0.3 (0.0–1.5) | 36; 0.3 (0.2–0.4) | 0; 0 | 7; 0.3 (0.1–0.7) | 0; 0 | 2; 0.1 (0.0–0.5) | 30; 0.2 (0.2–0.4) |
| Non-Hispanic Black or African American | 74; 20.3 (16.3–24.8) | 1903; 13.6 (13.1–14.2) | 4; 4.8 (1.3–11.7) | 247; 11.2 (9.9–12.5) | 5; 22.7 (7.8–45.4) | 158; 10.4 (8.9–12.0) | 1726; 14.3 (13.7–15.0) |
| Hispanic Black or African American | 2; 0.5 (0.1–2.0) | 34; 0.2 (0.2–0.3) | 0; 0 | 4; 0.2 (0.0–0.5) | 0; 0 | 1; 0.1 (0.0–0.4) | 32; 0.3 (0.2–0.4) |
| Other/unknown | 65; 17.9 (14.1–22.2) | 4059; 29.1 (28.3–29.8) | 21; 25.0 (16.2–35.6) | 581; 26.2 (24.4–28.1) | 4; 18.2 (5.2–40.3) | 444; 29.2 (27.0–31.6) | 3522; 29.3 (28.5–30.1) |
| Clinical characteristics | |||||||
| Liver fibrosis and cirrhosis | |||||||
| No evidence/history of liver fibrosis or cirrhosis | 318; 87.4 (83.5–90.6) | 13,093; 93.8 (93.3–94.2) | 55; 65.5 (54.3–75.5) | 1731; 78.2 (76.4–79.9) | 17; 77.3 (54.6–92.2) | 1138; 74.9 (72.7–77.1) | 11,389; 94.7 (94.3–95.1) |
| Any evidence of liver fibrosis | 2; 0.5 (0.1–2.0) | 99; 0.7 (0.6–0.9) | 1; 1.2 (0.0–6.5) | 88; 4.0 (3.2–4.9) | 1; 4.5 (0.1–22.8) | 72; 4.7 (3.7–5.9) | 68; 0.6 (0.4–0.7) |
| Any history of compensated cirrhosis | 18; 4.9 (3.0–7.7) | 373; 2.7 (2.4–3.0) | 10; 11.9 (5.9–20.8) | 220; 9.9 (8.7–11.3) | 2; 9.1 (1.1–29.2) | 175; 11.5 (10.0–13.2) | 267; 2.2 (2.0–2.5) |
| Any history of hepatic decompensation | 26; 7.1 (4.7–10.3) | 399; 2.9 (2.6–3.1) | 18; 21.4 (13.2–31.7) | 175; 7.9 (6.8–9.1) | 2; 9.1 (1.1–29.2) | 134; 8.8 (7.4–10.4) | 306; 2.5 (2.3–2.8) |
| NALD | |||||||
| No evidence | 350; 96.2 (93.6–97.9) | 13,454; 96.3 (96.0–96.7) | 82; 97.6 (91.7–99.7) | 2032; 91.8 (90.6–92.9) | 19; 86.4 (65.1–97.1) | 1391; 91.6 (90.1–92.9) | 11,591; 96.4 (96.0–96.7) |
| Any evidence | 14; 3.8 (2.1–6.4) | 510; 3.7 (3.3–4.0) | 2; 2.4 (0.3–8.3) | 182; 8.2 (7.1–9.4) | 3; 13.6 (2.9–34.9) | 128; 8.4 (7.1–9.9) | 439; 3.6 (3.3–4.0) |
| ALD | |||||||
| No evidence | 361; 99.2 (97.6–99.8) | 13,923; 99.7 (99.6–99.8) | 83; 98.8 (93.5–100.0) | 2205; 99.6 (99.2–99.8) | 22; 100 | 1517; 99.9 (99.5–100.0) | 11,996; 99.7 (99.6–99.8) |
| Any evidence | 3; 0.8 (0.2–2.4) | 41; 0.3 (0.2–0.4) | 1; 1.2 (0.0–6.5) | 9; 0.4 (0.2–0.8) | 0; 0 | 2; 0.1 (0.0–0.5) | 34; 0.3 (0.2–0.4) |
| Overweight/obesity | |||||||
| No evidence | 169; 46.4 (41.2–51.7) | 8105; 58.0 (57.2–58.9) | 33; 39.3 (28.8–50.5) | 1105; 49.9 (47.8–52.0) | 9; 40.9 (20.7–63.6) | 792; 52.1 (49.6–54.7) | 6955; 57.8 (56.9–58.7) |
| Any evidence – overweight | 110; 30.2 (25.5–35.2) | 3440; 24.6 (23.9–25.4) | 30; 35.7 (25.6–46.9) | 669; 30.2 (28.3–32.2) | 4; 18.2 (5.2–40.3) | 447; 29.4 (27.1–31.8) | 2962; 24.6 (23.9–25.4) |
| Any evidence – obesity | 85; 23.4 (19.1–28.0) | 2419; 17.3 (16.7–18.0) | 21; 25.0 (16.2–35.6) | 440; 19.9 (18.2–21.6) | 9; 40.9 (20.7–63.6) | 280; 18.4 (16.5–20.5) | 2113; 17.6 (16.9–18.3) |
| History of type 2 diabetes mellitus | |||||||
| No evidence | 318; 87.4 (83.5–90.6) | 12,922; 92.5 (92.1–93.0) | 70; 83.3 (73.6–90.6) | 1969; 88.9 (87.6–90.2) | 17; 77.3 (54.6–92.2) | 1357; 89.3 (87.7–90.8) | 11,155; 92.7 (92.2–93.2) |
| Any evidence | 46; 12.6 (9.4–16.5) | 1042; 7.5 (7.0–7.9) | 14; 16.7 (9.4–26.4) | 245; 11.1 (9.8–12.4) | 5; 22.7 (7.8–45.4) | 162; 10.7 (9.2–12.3) | 875; 7.3 (6.8–7.8) |
| Hyperlipidemia | |||||||
| No evidence | 313; 86.0 (82.0–89.4) | 11,807; 84.6 (83.9–85.1) | 67; 79.8 (69.6–87.7) | 1800; 81.3 (79.6–82.9) | 16; 72.7 (49.8–89.3) | 1227; 80.8 (78.7–82.7) | 10,155; 84.4 (83.8–85.1) |
| Any evidence | 51; 14.0 (10.6–18.0) | 2157; 15.4 (14.9–16.1) | 17; 20.2 (12.3–30.4) | 414; 18.7 (17.1–20.4) | 6; 27.3 (10.7–50.2) | 292; 19.2 (17.3–21.3) | 1875; 15.6 (14.9–16.2) |
| Hypertension | |||||||
| No evidence | 279; 76.6 (72.0–80.9) | 11,718; 83.9 (83.3–84.5) | 62; 73.8 (63.1–82.8) | 1768; 79.9 (78.1–81.5) | 12; 54.5 (32.2–75.6) | 1217; 80.1 (78.0–82.1) | 10,058; 83.6 (82.9–84.3) |
| Any evidence | 85; 23.4 (19.1–28.0) | 2246; 16.1 (15.5–16.7) | 22; 26.2 (17.2–36.9) | 446; 20.1 (18.5–21.9) | 10; 45.5 (24.4–67.8) | 302; 19.9 (17.9–22.0) | 1972; 16.4 (15.7–17.1) |
| Metabolic syndrome b | |||||||
| 0 | 136; 37.4 (32.4–42.6) | 6829; 48.9 (48.1–49.7) | 22; 26.2 (17.2–36.9) | 874; 39.5 (37.4–41.5) | 7; 31.8 (13.9–54.9) | 616; 40.6 (38.1–43.1) | 5876; 48.8 (47.9–49.7) |
| 1 | 132; 36.3 (31.3–41.4) | 4426; 31.7 (30.9–32.5) | 37; 44.0 (33.2–55.3) | 787; 35.5 (33.6–37.6) | 5; 22.7 (7.8–45.4) | 530; 34.9 (32.5–37.3) | 3777; 31.4 (30.6–32.2) |
| 2 | 59; 16.2 (12.6–20.4) | 1615; 11.6 (11.0–12.1) | 11; 13.1 (6.7–22.2) | 311; 14.0 (12.6–15.6) | 4; 18.2 (5.2–40.3) | 211; 13.9 (12.2–15.7) | 1436; 11.9 (11.4–12.5) |
| ≥3 | 37; 10.2 (7.3–13.7) | 1094; 7.8 (7.4–8.3) | 14; 16.7 (9.4–26.4) | 242; 10.9 (9.7–12.3) | 6; 27.3 (10.7–50.2) | 162; 10.7 (9.2–12.3) | 941; 7.8 (7.3–8.3) |
| Smoking | |||||||
| No evidence | 209; 57.4 (52.2–62.6) | 9206; 65.9 (65.1–66.7) | 38; 45.2 (34.3–56.5) | 1279; 57.8 (55.7–59.8) | 8; 36.4 (17.2–59.3) | 937; 61.7 (59.2–64.1) | 7920; 65.8 (65.0–66.7) |
| Any evidence | 155; 42.6 (37.4–47.8) | 4758; 34.1 (33.3–34.9) | 46; 54.8 (43.5–65.7) | 935; 42.2 (40.2–44.3) | 14; 63.6 (40.7–82.8) | 582; 38.3 (35.9–40.8) | 4110; 34.2 (33.3–35.0) |
| COPD | |||||||
| No evidence | 352; 96.7 (94.3–98.3) | 13,757; 98.5 (98.3–98.7) | 82; 97.6 (91.7–99.7) | 2179; 98.4 (97.8–98.9) | 21; 95.5 (77.2–99.9) | 1497; 98.6 (97.8–99.1) | 11,838; 98.4 (98.2–98.6) |
| Any evidence | 12; 3.3 (1.7,5.7) | 207; 1.5 (1.3–1.7) | 2; 2.4 (0.3–8.3) | 35; 1.6 (1.1–2.2) | 1; 4.5 (0.1–22.8) | 32; 2.1 (1.4–3.0) | 192; 1.6 (1.4–1.8) |
| Alcoholism | |||||||
| No evidence | 343; 94.2 (91.3–96.4) | 13,732; 98.3 (98.1–98.5) | 80; 95.2 (88.3–98.7) | 2166; 97.8 (97.1–98.4) | 20; 90.9 (70.8–98.9) | 1487; 97.9 (97.0–98.6) | 11,819; 98.2 (98.0–98.5) |
| Any evidence | 21; 5.8 (3.6–8.7) | 232; 1.7 (1.5–1.9) | 4; 4.8 (1.3–11.7) | 48; 2.2 (1.6–2.9) | 2; 9.1 (1.1–29.2) | 22; 1.4 (0.9–2.2) | 211; 1.8 (1.5–2.0) |
| Hospitalization c | |||||||
| No evidence | 289; 79.4 (74.9–83.4) | 13,351; 95.6 (95.3–95.9) | 51; 60.7 (49.5–71.2) | 2077; 93.8 (92.7–94.8) | 17; 77.3 (54.6–92.2) | 1442; 94.9 (93.7–96.0) | 11,465; 95.3 (94.9–95.7) |
| Any evidence | 75; 20.6 (16.6–25.1) | 613; 4.4 (4.1–4.7) | 33; 39.3 (28.8–50.5) | 137; 6.2 (5.2–7.3) | 5; 22.7 (7.8–45.4) | 77; 5.1 (4.0–6.3) | 565; 4.7 (4.3–5.1) |
| ALT flare | |||||||
| No history | 329; 90.4 (86.9–93.2) | 13,634; 97.6 (97.4–97.9) | 69; 82.1 (72.3–89.6) | 2098; 94.8 (93.7–95.7) | 16; 72.7 (49.8–89.3) | 1399; 92.1 (90.6–93.4) | 11,733; 97.5 (97.2–97.8) |
| Any history | 35; 9.6 (6.8–13.1) | 330; 2.4 (2.1–2.6) | – | – | – | – | 297; 2.5 (2.2–2.8) |
| Any history pre-treatment flare | – | – | 15; 17.9 (10.4–27.7) | 116; 5.2 (4.3–6.3) | 4; 18.2 (5.2–40.3) | 81; 5.3 (4.3–6.6) | – |
| Any history on-treatment flare | – | – | – | – | 2; 9.1 (1.1–29.2) | 21; 1.4 (0.9–2.1) | – |
| Any history pre- and on-treatment flare | – | – | – | – | 0 | 18; 1.2 (0.7–1.9) | – |
| CCI | |||||||
| N | 364; 100 | 13,964; 100 | 84; 100 | 2214; 100 | 22; 100 | 1519; 100 | 12,030; 100 |
| Mean (SD) | 2.1 (1.5) | 1.8 (1.1) | 2.7 (1.5) | 2.3 (1.0) | 2.8 (1.3) | 2.3 (1.0) | 1.8 (1.2) |
| Median (Q1–Q3) | 2.0 (2.0–2.0) | 2.0 (2.0–2.0) | 2.0 (2.0–3.0) | 2.0 (2.0–2.0) | 2.0 (2.0–4.0) | 2.0 (2.0–2.0) | 2.0 (2.0–2.0) |
| Virologic variables | |||||||
| HBV DNA | |||||||
| Undetectable | 51; 14.0 (10.6–18.0) | 2425; 17.4 (16.7–18.0) | 4; 4.8 (1.3–11.7) | 569; 25.7 (23.9–27.6) | 1; 4.5 (0.1–22.8) | 625; 41.1 (38.7–43.7) | 1996; 16.6 (15.9–17.3) |
| Detectable, viral load not available | 67; 18.4 (14.6–22.8) | 2365; 16.9 (16.3–17.6) | 15; 17.9 (10.4–27.7) | 242; 10.9 (9.7–12.3) | 6; 27.3 (10.7–50.2) | 226; 14.9 (13.1–16.8) | 2022; 16.8 (16.1–17.5) |
| Detectable, <2000 IU/mL | 80; 22.0 (17.8–26.6) | 5352; 38.3 (37.5–39.1) | 11; 13.1 (6.7–22.2) | 362; 16.4 (14.8–18.0) | 5; 22.7 (7.8–45.4) | 273; 18.0 (16.1–20.0) | 5024; 41.8 (40.9–42.6) |
| Detectable, 2000–<20,000 IU/mL | 39; 10.7 (7.7–14.4) | 1900; 13.6 (13.0–14.2) | 7; 8.3 (3.4–16.4) | 297; 13.4 (12.0–14.9) | 2; 9.1 (1.1–29.2) | 119; 7.8 (6.5–9.3) | 1639; 13.6 (13.0–14.3) |
| Detectable, ≥20,000 IU/mL | 127; 34.9 (30.0–40.0) | 1922; 13.8 (13.2–14.3) | 47; 56.0 (44.7–66.8) | 744; 33.6 (31.6–35.6) | 8; 36.4 (17.2–59.3) | 276; 18.2 (16.3–20.2) | 1349; 11.2 (10.7–11.8) |
ALD, alcoholic liver disease; ALT, alanine aminotransferase; CCI, Charlson Comorbidity Index; CI, confidence interval; COPD, chronic obstructive pulmonary disease; HBV, hepatitis B virus; IU, international units; NALD, non-alcoholic liver disease; Q, quartile; SD, standard deviation.
a95% CIs for proportions were calculated using the Clopper–Pearson (exact) method. 95% CIs are not presented for proportions equal to 0.0.
bA count variable indicating the presence of the four components of metabolic syndrome: overweight/obesity, type 2 diabetes mellitus, hypertension, and hyperlipidemia.
cHospitalization is defined as hospital overnight stay (i.e., discharge date > admission date) due to any cause and recorded on or in the 3 months before the respective index date.
Flare patients were more likely than no-flare patients to have evidence/history of liver fibrosis or cirrhosis (12.6% [95% CI 9.4–16.5%] vs 6.2% [95% CI 5.8–6.7%]); specifically, flare patients were more likely to have a history of compensated cirrhosis (4.9% [95% CI 3.0–7.7%] vs 2.7% [95% CI 2.4–3.0%]) or a history of hepatic decompensation (7.1% [95% CI 4.7–10.3%] vs 2.9% [95% CI 2.6–3.1%]) (Table 2), and were also more likely to have evidence of advanced liver disease (19.2% [95% CI 15.3–23.7%] vs 4.1% [95% CI 3.8–4.5%]), as defined by APRI or FIB-4 score (Supplemental Table 6). Other imbalances between flare and no-flare patients included evidence of being overweight (30.2% [95% CI 25.5–35.2%] vs 24.6% [95% CI 23.9–25.4%]) or of obesity (23.4% [95% CI 19.1–28.0%] vs 17.3% [95% CI 16.7–18.0%]), evidence of hypertension (23.4% [95% CI 19.1–28.0%] vs 16.1% [15.5–16.7%]), evidence of smoking (42.6% [95% CI 37.4–47.8%] vs 34.1% [95% CI 33.3–34.9%]), evidence of alcoholism (5.8% [95% CI 3.6–8.7%] vs 1.7% [95% CI 1.5–1.9%]), evidence of hospitalization (20.6% [95% CI 16.6–25.1%] vs 4.4% [95% CI 4.1–4.7%]) (Table 2), and a history of ALT flares (9.6% [95% CI 6.8–13.1%] vs 2.4% [95% CI 2.1–2.6%]) (Supplemental Table 6).
Patients who experienced an ALT flare were more likely to be HBeAg-positive than no-flare patients (37.7% [95% CI 31.2–44.5%] vs 14.5% [95% CI 13.8–15.3%]) and, while all patients were not HBsAg-negative by design, flare patients were less likely to have missing HBsAg data (25.8% vs 31.9%) (Supplemental Table 6). Baseline ALT levels were higher in flare than in no-flare patients; a higher proportion had levels >2 × ULN (35.4% [95% CI 30.5–40.6%] vs 6.4% [95% CI 6.0–6.8%]) and >5 × ULN (15.1% [95% CI 11.6–19.2%] vs 0.9% [95% CI 0.8–1.1%]). More flare patients than no-flare patients also had HBV DNA levels ≥20,000 IU/mL (34.9% [95% CI 30.0–40.0%] vs 13.8% [95% CI 13.2–14.3%]) (Table 2).
Treatment initiation cohort
The differences between flare and no-flare patients observed in the untreated cohort were generally mirrored in the treatment initiation cohort (Table 2; Supplemental Table 6); the mean age at treatment initiation index date was similar (46.3 years and 48.3 years), and there were higher proportions of males (73.8% [95% CI 63.1–82.8%] vs 58.6% [95% CI 56.5–60.6%]) and non-Hispanic Caucasians (32.1% [95% CI 22.4–43.2%] vs 14.8% [95% CI 13.3–16.3%]).
Differences between flare and no-flare patients were again observed for some evidence/history of fibrosis or cirrhosis (34.5% [95% CI 24.5–45.7%] vs 21.8% [95% CI 20.1–23.6%]), history of hepatic decompensation (21.4% [95% CI 13.2–31.7%] vs 7.9% [95% CI 6.8–9.1%]), evidence of advanced liver disease (39.3% [95% CI 28.8–50.5%] vs 8.7% [95% CI 7.5–9.9%]), evidence of hospitalization (39.3% [95% CI 28.8–50.5%] vs 6.2% [95% CI 5.2–7.3%]), and a history of ALT flares (17.9% [95% CI 10.4–27.7%] vs 5.2% [95% CI 4.3–6.3%]) (Table 2; Supplemental Table 6). Patients who experienced an ALT flare were also more likely to be HBeAg-positive than no-flare patients (48.5% [95% CI 36.0–61.1%] vs 27.9% [95% CI 25.7–30.1%]); differences in other biochemical and virologic parameters described above for the untreated cohort persisted in the treatment initiation cohort.
Most patients (98.6%) initiated NA monotherapy; TDF (53.5%) was the most commonly prescribed NA, followed by ETV (31.0%) and TAF (14.1%). Only 1.2% of patients initiated NA-NA combination therapy, and very few patients (0.2%) initiated IFN monotherapy. Few differences were observed between flare and no-flare patients with regard to anti-HBV therapies, although within those initiating NA therapy, more flare (65.5%) than no-flare patients (53.1%) initiated TDF.
Treatment discontinuation cohort
Too few patients (n = 22, 1.4%) in this cohort experienced a flare to warrant comparison with no-flare patients. Among those who did experience flares, 22.7% had some evidence/history of fibrosis or cirrhosis, and 18.2% had evidence of advanced liver disease (Table 2; Supplemental Table 6) at the treatment discontinuation index date. Evidence of obesity, hypertension, and hospitalization was present in 40.9%, 45.5%, and 22.7% of flare patients, respectively (Table 2). More than a quarter of flare patients (27.3%) had a history of ALT flares, and 36.4% had HBV DNA levels ≥20,000 IU/mL (Supplemental Table 6; Table 2).
Risk factors for ALT flares
In the untreated cohort, flare patients were more likely than no-flare patients to be male and of younger age (18–<30 years) (Table 3A). Flare patients were also more likely than no-flare patients to have a history of ALT flares, evidence of liver fibrosis, history of compensated cirrhosis, history of hepatic decompensation, or evidence of metabolic syndrome.
Table 3.
Analysis of risk factors for ALT flares.
| Pre-treatment ALT flares | |||
|---|---|---|---|
| HR (95% CI) | p-value | ||
| (A) Untreated cohort | |||
| Sex | Female | Ref | |
| Male | 1.80 (1.43–2.26) | <0.001 * | |
| Age at study entry index date | 18–<30 years | Ref | |
| 30–<40 years | 0.70 (0.48–1.04) | 0.077 | |
| 40–<50 years | 0.81 (0.56–1.18) | 0.275 | |
| 50–<60 years | 0.67 (0.45–0.99) | 0.047 * | |
| 60–<70 years | 0.72 (0.47–1.12) | 0.142 | |
| ≥70 years | 0.42 (0.21–0.84) | 0.015 * | |
| Race | White | Ref | |
| Black | 1.03 (0.75–1.41) | 0.844 | |
| Asian | 0.81 (0.61–1.07) | 0.144 | |
| Other | 0.58 (0.41–0.81) | 0.001 * | |
| History of ALT flares | No history of ALT flares | Ref | |
| History of ALT flares | 2.84 (1.97–4.09) | <0.001 * | |
| Liver fibrosis or cirrhosis | No evidence/history of liver fibrosis or cirrhosis | Ref | |
| Any evidence of liver fibrosis | 2.40 (1.31–4.40) | 0.005 * | |
| Any history of compensated cirrhosis | 1.73 (1.10–2.73) | 0.017 * | |
| Any history of hepatic decompensation | 3.75 (2.75–5.13) | <0.001 * | |
| Metabolic syndrome | Score=0 | Ref | |
| Score ≥1 | 1.51 (1.17–1.95) | 0.002 * | |
| Smoking/COPD | No evidence of smoking or COPD | Ref | |
| Evidence of smoking or COPD | 1.13 (0.91–1.42) | 0.272 | |
ALT, alanine aminotransferase; CI, confidence interval; COPD, chronic obstructive pulmonary disease; HR, hazard ratio; Ref, Reference.
*Statistical significance at the 0.05 level.
In the treatment initiation cohort, flare patients were more likely than no-flare patients to be of younger age (18–<40 years), of Black or Asian race (compared with White), have a history of ALT flares, or have evidence of liver fibrosis, cirrhosis, or hepatic decompensation (Table 3B).
Characteristics of ALT flares
Among patients experiencing an ALT flare within each cohort, the mean time to first flare from the respective index dates was 635.1 days in the untreated cohort, 200.4 days in the treatment initiation cohort, and 249.0 days in the treatment discontinuation cohort. Median time to ALT flare was not reached in any of the cohorts. The mean duration of the first flare was similar across the three groups (108.5 days, 94.0 days, and 94.7 days, respectively). The mean peak value of the flares was 652.5 IU/mL in the untreated cohort, 565.5 IU/mL in the treatment initiation cohort, and 516.8 IU/mL in the treatment discontinuation cohort. There was no difference in the proportion of patients experiencing first flare peaks >20 × ULN (untreated 21.4%, 95% CI 17.3–26.0%; treatment initiation 17.9%, 95% CI 10.4–27.7%; treatment discontinuation 9.1%, 95% CI 1.1–29.2%) (Table 1; Supplemental Figure 1).
Changes in ALT levels versus other laboratory parameters over time among patients with ALT flares
In all cohorts, HBV DNA levels increased from 6 to 9 months before the ALT peak (Supplemental Figures 2–4); peak levels occurred before the ALT peak in the untreated (approximately 1 month) and treatment initiation (approximately 4 months) cohorts, and corresponded with the ALT peak in the treatment discontinuation cohort. Correlations in the trend of serum bilirubin versus ALT values around the peak of ALT flares were similar across all three cohorts (Supplemental Figures 2–4). Serum bilirubin increase aligned with ALT flares, but peaked approximately 1 month after the ALT peak. In the untreated and treatment initiation cohorts, serum bilirubin at the ALT peak was approximately 8–9 × ULN (150–170 μmol/L), assuming a ULN definition of 18.81 μmol/L. 24 Variations in other laboratory parameters are described in the Supplemental Material.
Associations between ALT flares and outcomes
Following adjustment for confounding and non-informative censoring, untreated patients who experienced ALT flares were at increased risk of hepatic decompensation (aOR [95% CI] 2.7 [1.7–4.2]), hospitalization (aOR [95% CI] 1.5 [1.0–2.2]), or death (aOR [95% CI] 3.8 [2.4–6.2]). They were also more likely to experience spontaneous HBsAg loss (aOR [95% CI] 2.4 [1.2–5.1]) (Figure 3(A)).
Figure 3.
Forest plot of adjusted associations between clinical and virologic outcomes and ALT flares in the untreated cohort (A) and treatment initiation cohort (B).
In the treatment initiation cohort, ALT flares were not associated with an increased risk of hepatic decompensation (aOR [95% CI] 1.3 [0.5–3.7]), hospitalization (aOR [95% CI] 0.7 [0.3–1.6]), or death (aOR [95% CI] 1.3 [0.5–3.4]). Moreover, in addition to increasing the likelihood of HBsAg loss fifteen-fold (aOR [95% CI] 15.3 [4.5–51.8]), patients with ALT flares were more likely to experience HBeAg loss (aOR [95% CI] 3.7 [1.9–7.3) than no-flare patients (Figure 3(B)).
Discussion
The dynamic role of ALT flares during the course of HBV infection has been previously described, emphasizing the importance of investigating ALT flares in different patient populations and different scenarios to inform optimal management. 10 In this US-based, retrospective cohort study using EHR data among a diverse population of patients with chronic HBV infection, the IR of ALT flares in the cohort of patients initiating treatment was markedly higher than in the untreated and treatment discontinuation cohorts. We also found that the occurrence of ALT flares in the untreated group increased the risk of adverse clinical outcomes and loss of HBeAg and HBsAg, while the occurrence of flares in the cohort of patients initiating mostly NA treatment increased the likelihood of HBsAg and HBeAg loss but not of the adverse clinical outcomes investigated.
The difference in IRs of ALT flares across the study cohorts may indicate varying underlying biologic mechanisms of flares in different phases of the disease. Flares can occur during the natural course of infection and may indicate a transition from active to inactive disease. 10 Such flares are usually self-limiting and associated with a decrease in serum HBV DNA level, loss of HBeAg, and development of anti-HBe and anti-HBs seroconversion. 10 Other ALT flares can, however, be prolonged and associated with progressive liver injury, resulting in hepatic decompensation or death. 10 The incidence of ALT flares can be influenced by both the host and viral factors. Age is an important host factor, with flares more common in younger versus older adults, and some studies have reported flares to be more common in men than in women.25,26 Consistent with this, among the untreated group in the present study, a higher proportion of men than women was noted in those who experienced ALT flares. In addition, older age was associated with a lower risk of flares. We also observed higher proportions of patients at baseline who were HBeAg-positive and had HBV DNA ≥20,000 IU/mL among untreated patients who experienced ALT flares; this also aligns with previous reports. 26
ALT flares following initiation of treatment can occur with both IFNs and NAs but are more common with IFNs. 10 Among patients treated with NAs, ALT flares are more common in HBeAg-positive than in HBeAg-negative patients10,27–29; development of these flares is probably linked to an enhanced antiviral T-cell response.10,30 These flares are considered beneficial as they may contribute to a higher rate of HBeAg seroconversion in the first year of NA treatment than in subsequent years of treatment, while hepatic decompensation is uncommon. 10 The vast majority of patients initiating treatment in the current study were prescribed NAs (TDF or ETV). The mean time to onset of flares was 200 days, and the proportion of patients who were HBeAg-positive at baseline was greater in those who experienced flares than in those who did not. The late onset of flares in the treatment initiation cohort, along with the prior increase in HBV DNA, may indicate a degree of non-adherence to treatment, although this was not evaluated in the current study and hence is purely hypothetical.
As indicated previously, the occurrence of ALT flares in the untreated cohort increased the risk of adverse clinical outcomes (i.e., hepatic decompensation, all-cause hospitalization, and all-cause mortality [most likely due to disease progression]), as well as HBsAg loss and HBeAg loss. These findings are consistent with previous reports.24,31 In the cohort of patients initiating treatment (mostly NAs), ALT flares increased the likelihood of HBeAg loss and, interestingly, HBsAg loss. Quantitative HBsAg testing is not performed routinely in the US, hence serial HBsAg levels were not available in the current study to provide further insights. However, future research to shed more light on the relationship between ALT flares and HBsAg loss would be of interest. While ALT flares in the treatment initiation cohort did not increase the risk of adverse clinical outcomes such as hepatic decompensation and all-cause mortality, we cannot exclude the possibility of a lack of power to show an association due to insufficient sample size. In addition, while marginal structural models were used to estimate the associations between ALT flares and the outcomes of interest, residual confounding may persist and we cannot exclude the possibility that ALT flares represent markers of advanced disease rather than causal drivers of decompensation.
Among patients who initiated treatment in the current study, approximately 75% discontinued treatment; this is higher than reported in another recent claims-based study from the US (20%). 32 The reasons for discontinuation were not captured for this study. One possible factor is that the definition of discontinuation used was based (at least in part) on the end date of the last continuous prescription, which may not always correlate with the end of supply. It is also possible that a proportion of the discontinuation cohort were non-adherent to their treatment, given the high HBV DNA levels observed in some patients at the treatment discontinuation index date.
Our study has several strengths, notably the inclusion of a large (>14,000), ethnically diverse overall cohort, the use of marginal structural models to account for time-dependent confounding, and evaluation of both virological and clinical outcomes. Some possible limitations should also be considered when interpreting study data. Firstly, the Optum EHR database only provided data from 2007 onwards, potentially introducing left truncation bias. To mitigate this, only patients with first evidence of chronic HBV infection since 2012 were included, allowing for clinical history assessments. Secondly, biochemical hepatic decompensation may arise from other causes (such as renal or multi-organ failure), and so patients not diagnosed with hepatic decompensation may be subject to misclassification. There is also the possibility of misclassification bias and under-reporting due to irregular data recording and a lag in mortality reporting, as well as incomplete patient information if patients switched healthcare providers (although data suggest good persistence within Optum’s network). Further possible limitations include the lack of data on HBV genotype and treatment adherence, and possible misclassification of ALT flares due to irregular testing intervals. Finally, although covariates were adjusted for in the analysis, it is possible that unmeasured confounders may have influenced the results.
The definition of ALT flare used may influence the observed incidence. We employed a definition of ALT >5x ULN and >2x baseline or nadir ALT level; an abrupt increase in serum ALT 5x ULN is the minimum requirement among the most common definitions,9–11 while ALT >2x baseline or nadir was added to rule out possible transient fluctuations in ALT levels. A more stringent definition (based on ALT >10x ULN) was considered; however, the number of patients across the different groups experiencing an ALT flare by this definition was very small (data not shown), precluding further modeling analysis. Similarly and conversely, ALT elevations in chronic HBV infection can result from several non-HBV causes, including alcohol use, fatty liver disease, and drug-induced liver injury. The potential exists, therefore, for exposure misclassification that may have non-differentially biased findings to the null. In addition, we used a fixed ALT ULN of 40 U/L in order to standardize flare identification across laboratories and health systems within the EHR database, consistent with historical laboratory reference standards and prior HBV epidemiology studies. 33 However, we acknowledge that this may over-estimate normal values, particularly for women.
Conclusions
ALT flares in untreated patients with chronic HBV infection were associated with both increased HBsAg loss and higher risk of adverse clinical outcomes. In contrast, flares occurring after treatment initiation were associated with improved virologic responses without increased risk of clinical complications. Several risk factors for ALT flares were common in the untreated group and in those initiating treatment (including age, a history of ALT flares, and evidence/history of liver fibrosis, compensated cirrhosis, or hepatic decompensation). Greater understanding of the risk factors and consequences of ALT flares in patients with chronic HBV infection is important for providing optimal care.
Supplemental material
Supplemental Material for Association of alanine aminotransferase flares with hepatitis B surface antigen loss and clinical outcomes in treated and untreated patients with chronic hepatitis B virus infection: A US retrospective cohort study by Myriam Drysdale, Eleonora Morais, Rose Chang, Shuang Wang, Megan Pinaire, Manasi Mohan, Rui Song, Mei Shung Duh, Grace Dolman, Dickens Theodore and Iain A. Gillespie in Therapeutic Advances in Infectious Disease.
Acknowledgements
Editorial support (in the form of writing assistance, including preparation of the draft manuscript under the direction and guidance of the authors, collating and incorporating authors’ comments for each draft, assembling tables, grammatical editing, and referencing) was provided by Vijayalatha Venugopalan, PhD, and Tony Reardon, PhD, of Luna, OPEN Health Communications, in accordance with Good Publication Practice (GPP) guidelines (https://www.ismpp.org/gpp-2022). The support was funded by GSK.
Author note: All named authors meet the International Committee of Medical Journal Editors criteria for authorship of this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.
Author contributions: All authors made a significant contribution to the work reported, whether that was in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising, or critically reviewing the manuscript; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Funding: The authors disclosed receipt of the following financial support for the research, and/or publication of this article: This study was funded by GSK (study number 212764).
The authors declared the following potential conflicts of interest with respect to the research, and/or publication of this article: Authors MD, EM, GD, DT, and IAG are employees of and/or hold financial equities in GSK. RC, SW, MP, MM, RS, and MSD are employees of Analysis Group Inc.; Analysis Group Inc. received funding from GSK to conduct this study.
Supplemental material: Supplemental material for this article is available online.
ORCID iD
Myriam Drysdale https://orcid.org/0000-0002-8994-2816
Ethical considerations
This observational study was performed in accordance with ethical principles that are consistent with the Declaration of Helsinki and its amendments, the Guidelines for Good Pharmacoepidemiology Practice, and the applicable legislation on non-interventional studies and/or observational studies. This study complied with all applicable laws regarding subject privacy. No direct subject contact or primary collection of individual human subject data occurred. Study results are in tabular form and presented as aggregate analyses that omit subject identification; therefore, informed consent and ethics committee or institutional review board approval were not required. Any publications and reports do not include subject identifiers. Data from the Optum EHR dataset were de-identified and compliant with the patient requirements of the Health Insurance Portability and Accountability Act.
Data availability statement
Data for this study are not publicly available but may be obtained on request to the corresponding author (myriam.g.drysdale@gsk.com).
References
- 1.Polaris Observatory Collaborators . Global prevalence, cascade of care, and prophylaxis coverage of hepatitis B in 2022: a modelling study. Lancet Gastroenterol Hepatol 2023; 8: 879–907. 10.1016/S2468-1253(23)00197-8 [DOI] [PubMed] [Google Scholar]
- 2.Marcellin P, Wong DK, Sievert W, et al. Ten-year efficacy and safety of tenofovir disoproxil fumarate treatment for chronic hepatitis B virus infection. Liver Int 2019; 39: 1868–1875. 10.1111/liv.14155 [DOI] [PubMed] [Google Scholar]
- 3.Hsu YC, Huang DQ, Nguyen MH. Global burden of hepatitis B virus: current status, missed opportunities and a call for action. Nat Rev Gastroenterol Hepatol 2023; 20: 524–537. 10.1038/s41575-023-00760-9 [DOI] [PubMed] [Google Scholar]
- 4.Lim JK, Nguyen MH, Kim WR, et al. Prevalence of chronic hepatitis B virus infection in the United States. Am J Gastroenterol 2020; 115: 1429–1438. 10.14309/ajg.0000000000000651 [DOI] [PubMed] [Google Scholar]
- 5.World Health Organization . Global hepatitis report 2024: action for access in low- and middle-income countries. 2024; (accessed November 2025). https://www.who.int/publications/i/item/9789240091672 [Google Scholar]
- 6.Centers for Disease Control and Prevention . 2023 Viral Hepatitis Surveillance Report. 2025; (accessed November 2025). https://www.cdc.gov/hepatitis-surveillance-2023/about/index.html [Google Scholar]
- 7.GBD 2019 Hepatitis B Collaborators . Global, regional, and national burden of hepatitis B, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Gastroenterol Hepatol 2022; 7: 796–829. 10.1016/S2468-1253(22)00124-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fontana RJ, Avigan MI, Janssen HLA, et al. Liver safety assessment in clinical trials of new agents for chronic hepatitis B. J Viral Hepat 2020; 27: 96–109. 10.1111/jvh.13223 [DOI] [PubMed] [Google Scholar]
- 9.Choi HSJ, Sonneveld MJ, Farag MS, et al. Effects of on-treatment ALT flares on serum HBsAg and HBV RNA in patients with chronic HBV infection. J Viral Hepat 2021; 28: 1729–1737. 10.1111/jvh.13613 [DOI] [PubMed] [Google Scholar]
- 10.Ghany MG, Feld JJ, Chang KM, et al. Serum alanine aminotransferase flares in chronic hepatitis B infection: the good and the bad. Lancet Gastroenterol Hepatol 2020; 5: 406–417. 10.1016/S2468-1253(19)30344-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jeng WJ, Chen YC, Liaw YF. Great and rapid HBsAg decline in patients with on-treatment hepatitis flare in early phase of potent antiviral therapy. J Viral Hepat 2018; 25: 421–428. 10.1111/jvh.12833 [DOI] [PubMed] [Google Scholar]
- 12.European Association for the Study of the Liver . EASL Clinical Practice Guidelines on the management of hepatitis B virus infection. J Hepatol 2025; 83: 502–583. 10.1016/j.jhep.2025.03.018 [DOI] [PubMed] [Google Scholar]
- 13.Terrault NA, Lok ASF, McMahon BJ, et al. Update on prevention, diagnosis, and treatment of chronic hepatitis B: AASLD 2018 hepatitis B guidance. Hepatology 2018; 67: 1560–1599. 10.1002/hep.29800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sarin SK, Kumar M, Lau GK, et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: a 2015 update. Hepatol Int 2016; 10: 1–98. 10.1007/s12072-015-9675-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lok ASF. Toward a functional cure for hepatitis B. Gut Liver 2024; 18: 593–601. 10.5009/gnl240023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ren H, Huang Y. Effects of pegylated interferon-α based therapies on functional cure and the risk of hepatocellular carcinoma development in patients with chronic hepatitis B. J Viral Hepat 2019; 26(Suppl 1): 5–31. 10.1111/jvh.13150 [DOI] [PubMed] [Google Scholar]
- 17.Slaets L, De Ridder F, Lenz O, et al. Systematic review with meta-analysis: hepatitis B surface antigen decline and seroclearance in chronic hepatitis B patients on nucleos(t)ide analogues or pegylated interferon therapy. Gastroenterol Hepatol 2020; 2: 106–116. 10.1002/ygh2.393 [DOI] [Google Scholar]
- 18.Dongelmans EJ, Hirode G, Hansen BE, et al. Predictors of hepatic flares after nucleos(t)ide analogue cessation – results of a global cohort study (RETRACT-B study). J Hepatol 2025; 82: 446–455. 10.1016/j.jhep.2024.08.015 [DOI] [PubMed] [Google Scholar]
- 19.Du Y, Du B, Fang X, et al. ALT flare predicts hepatocellular carcinoma among antiviral treated patients with chronic hepatitis B: a cross-country cohort study. Front Oncol 2021; 10: 615203. 10.3389/fonc.2020.615203 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chang ML, Cheng JS, Chien RN, et al. Hepatitis flares are associated with better outcomes than no flare in patients with decompensated cirrhosis and chronic hepatitis B virus infection. Clin Gastroenterol Hepatol 2020; 18: 2064–2072.e2. 10.1016/j.cgh.2020.01.018 [DOI] [PubMed] [Google Scholar]
- 21.Yao N, Wang J, Wu Y, et al. Frequent alanine aminotransferase flares and promising antiviral therapy efficacy during the preschool age: an opportunity to achieve HBsAg loss in children with mother-to-child transmitted hepatitis B. J Viral Hepat 2022; 29: 748–755. 10.1111/jvh.13720 [DOI] [PubMed] [Google Scholar]
- 22.von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61: 344–349. 10.1016/j.jclinepi.2007.11.008 [DOI] [PubMed] [Google Scholar]
- 23.Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 2008; 168: 656–664. 10.1093/aje/kwn164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cremer J, Elston R, Campbell FM, et al. B-Clear phase 2b study design: establishing the efficacy and safety of bepirovirsen in patients with chronic hepatitis B virus infection. Adv Ther 2023; 40: 4101–4110. 10.1007/s12325-023-02531-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Brahmania M, Lombardero M, Hansen BE, et al. Association between severe serum alanine aminotransferase flares and hepatitis B e antigen seroconversion and HBV DNA decrease in untreated patients with chronic HBV infection. Clin Gastroenterol Hepatol 2019; 17: 2541–2551.e2. 10.1016/j.cgh.2019.02.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lok AS, Lai CL. Acute exacerbations in Chinese patients with chronic hepatitis B virus (HBV) infection. Incidence, predisposing factors and etiology. J Hepatol 1990; 10: 29–34. 10.1016/0168-8278(90)90069-4 [DOI] [PubMed] [Google Scholar]
- 27.Chang TT, Gish RG, de Man R, et al. A comparison of entecavir and lamivudine for HBeAg-positive chronic hepatitis B. N Engl J Med 2006; 354: 1001–1010. 10.1056/nejmoa051285 [DOI] [PubMed] [Google Scholar]
- 28.Lai CL, Shouval D, Lok AS, et al. Entecavir versus lamivudine for patients with HBeAg-negative chronic hepatitis B. N Engl J Med 2006; 354: 1011–1020. 10.1056/NEJMoa051287 [DOI] [PubMed] [Google Scholar]
- 29.Marcellin P, Heathcote EJ, Buti M, et al. Tenofovir disoproxil fumarate versus adefovir dipivoxil for chronic hepatitis B. N Engl J Med 2008; 359: 2442–2455. 10.1056/NEJMoa0802878 [DOI] [PubMed] [Google Scholar]
- 30.Boni C, Penna A, Bertoletti A, et al. Transient restoration of anti-viral T cell responses induced by lamivudine therapy in chronic hepatitis B. J Hepatol 2003; 39: 595–605. 10.1016/s0168-8278(03)00292-7 [DOI] [PubMed] [Google Scholar]
- 31.Chang ML, Liaw YF. Hepatitis B flares in chronic hepatitis B: pathogenesis, natural course, and management. J Hepatol 2014; 61: 1407–1417. 10.1016/j.jhep.2014.08.033 [DOI] [PubMed] [Google Scholar]
- 32.Anderson S, Gielen V, Coutinho AD, et al. Duration of nucleos(t)ide analogue treatments in patients with chronic hepatitis B virus infection in the United States. J Viral Hepat 2025; 32: e70055. 10.1111/jvh.70055 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Neuschwander-Tetri BA, Unlap A, Creer MH. The upper limits of normal for serum ALT levels reported by clinical laboratories depend on local reference populations. Arch Intern Med 2004; 168: 663–666. [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
Supplemental Material for Association of alanine aminotransferase flares with hepatitis B surface antigen loss and clinical outcomes in treated and untreated patients with chronic hepatitis B virus infection: A US retrospective cohort study by Myriam Drysdale, Eleonora Morais, Rose Chang, Shuang Wang, Megan Pinaire, Manasi Mohan, Rui Song, Mei Shung Duh, Grace Dolman, Dickens Theodore and Iain A. Gillespie in Therapeutic Advances in Infectious Disease.
Data Availability Statement
Data for this study are not publicly available but may be obtained on request to the corresponding author (myriam.g.drysdale@gsk.com).