|Home | About | Journals | Submit | Contact Us | Français|
The diagnosis of acute hepatitis C virus (HCV) infection is imprecise because antibody testing does not differentiate between acute and chronic infection. While virologic features, such as viral load fluctuations and low levels of viremia, have been noted to be characteristic of acute HCV infection, these parameters have not been utilized for diagnosis.
We validated the use of these novel parameters (i.e., viral load fluctuations>1 log and HCV-RNA<100,000 IU/mL) in a cohort of acute HCV seroconverters. We then applied standard diagnostic criteria for acute HCV in a cohort of high-risk injection drug users entering prison with suspected acute HCV infection (n=37). We subsequently assessed whether these novel virologic parameters, measured serially over a 10-week period, could enhance the diagnosis of acute infection.
Low-level viremia and viral fluctuations were highly prevalent in our cohort of acute seroconverters (81 and 86 percent, respectively) while low-level viremia occurred in only 13% of control patients with chronic infection. Using standard criteria, 37 inmates were diagnosed with acute HCV infection. In the 35 patients with HCV-RNA at baseline, we found low-level viremia to be highly prevalent (n=27; 77%); in patients with a minimum of two HCV-RNA samples, we demonstrated viral fluctuations in more than one-third (n=9; 36%).
The diagnosis of acute infection in the HCV-seropositive patient is strengthened by the use of virologic parameters that are uncommon in chronic disease. Viral load fluctuations and low levels of HCV-RNA should be incorporated into standard diagnostic criteria.
The diagnosis of acute hepatitis C virus (HCV) infection is problematic because antibody testing does not differentiate between acute and chronic infection and seroconversion is often not documented [1, 2]. Physicians have traditionally made clinical judgments based on the presence of symptoms, viremia, abnormal aminotransferases, and a history of prior seronegativity . Accurate identification of the acutely-infected individual is important since early antiviral therapy can lead to higher sustained virologic response (SVR) rates [4–6].
Many persons newly-infected with HCV are asymptomatic and do not present for medical evaluation. The diagnosis of HCV infection includes arbitrary thresholds for defining severe hepatitis; yet patients may be identified either early or late in the course of their acute infection when aminotransferases may be minimally abnormal . In this clinical scenario, the diagnosis can be missed or mistaken for chronic infection. Conversely, aminotransferases may be markedly elevated due to a variety of other causes, including alcohol abuse, drug toxicity, or cocaine intoxication [8, 9]; in any of these scenarios, a patient with chronic HCV infection and elevated aminotransferases may be misclassified with newly-acquired infection. Thus, symptoms, signs, and laboratory testing all have their inherent limitations in clinical practice.
Some individuals are diagnosed due to their history of risk. This is particularly true of healthcare workers (HCWs) who report a precisely defined exposure with subsequent confirmation of seroconversion . Although they account for a small portion of acute infections, HCWs are often well-represented in studies of acute HCV due to their prospective identification . Injection drug users (IDUs) often acquire HCV infection within 6–12 months of initiation of sharing drug paraphernalia, [11, 12], but rarely seek medical attention [4, 5, 13, 14]. They are also unlikely to have documentation of prior serologic testing due to lack of health-seeking behaviors and fragmented care [15, 16]. Since the median time from initial viremia to seroconversion is approximately one month , the majority of patients are already seropositive during the early stages of infection.
Without knowledge of prior serologic testing, differentiating acute from chronic infection in IDUs is difficult; therefore, other distinguishing features are needed to improve diagnostic accuracy. Chronically-infected patients have stable HCV-RNA levels that vary by approximately 0.5 log [17–20]. An individual’s viral “set-point” is maintained over time in the absence of other modifying factors (eg, concomitant HIV infection) . In contrast, in acute HCV infection, viral-host interactions lead to marked variations in viremia until the ultimate clinical outcome (i.e., spontaneous clearance or viral persistence) is established [22–24].
Another helpful diagnostic parameter is the magnitude of the HCV-RNA levels. In a study of 2,472 chronically-infected patients, only 8.9 percent had a viral load <5.0 log10IU/ml . Another retrospective study of 1550 patients with chronic infection demonstrated that only 13 percent had a viral load <400,000 IU/mL . In contrast, studies of acute HCV infection have demonstrated that many patients have low levels of viremia, which may reflect either viral decline immediately prior to spontaneous clearance or transient immunologic control [14, 23, 27–33].
Although viral load fluctuations and low-level viremia have been noted in multiple reports of acute HCV infection, these parameters have never been applied as part of a diagnostic approach. Within this report, we provide data on 37 individuals with recent onset of IDU who were diagnosed with acute/early HCV infection shortly after being incarcerated, using criteria endorsed by the Centers for Disease Control (CDC). We demonstrate that viral load fluctuations of >1 log and/or low levels of viremia <105 IU/mL were highly prevalent in these patients. We also show that these virologic parameters were highly prevalent in a cohort of patients with documented seroconversion; in contrast, low-level viremia was distinctly uncommon in our controls with chronic HCV. We propose that the diagnosis of acute HCV infection can be well-determined by initial risk ascertainment followed by serial changes in viral load. These parameters are particularly helpful in assessing patients without prior serologic testing.
In our prison-based pilot study, we identified 21 recently-incarcerated inmates with acute HCV infection during a 30-month period based on referrals for symptoms of hepatitis and/or elevated aminotransferases . We demonstrated that serial monitoring of HCV-RNA at baseline, four and 10 weeks differentiated patients who developed persistent viremia from those who attained spontaneous resolution.
Our subsequent study (the basis of this report) was designed to identify acute HCV cases based on high-risk behaviors through systematic screening of 3,248 newly-incarcerated inmates (October 2006-December 2007). Questionnaires were administered upon intake by healthcare providers screening for a history of HCV testing and new onset IDU or sharing of paraphernalia within 12 months of incarceration. Fifty-four high-risk patients without a self-reported history of HCV infection were screened serologically and invited to enroll in an observational study.
The duration of infection was calculated using the number of months between estimated inoculation and enrollment. Laboratories included: HCV-RNA (Versant HCVRNAv3.0 assay bDNA), aminotransferases, and hepatitis A and B serologies. Regardless of whether aminotransferases were abnormal, patients underwent screening for HCV-RNA and aminotransferases at baseline, four, and ten weeks based on risk. Those with confirmed hepatitis were seen at the subspecialty clinic for medical evaluation.
We adopted two approaches to the diagnosis of acute hepatitis C. Most studies of acute HCV have required prior HCV antibody serologies within six months of presentation, detectable HCV-RNA and abnormal aminotransferases, ranging from 2–20x the upper limit of normal (ULN) . For our “standard approach”, we utilized an alanine aminotransferase (ALT) level >7xULN as our diagnostic threshold, as defined by the CDC . Patients with recent onset of high-risk behaviors were categorized in terms of probability of acute HCV infection as follows: a) seronegative patients with HCV-RNA at baseline and subsequent seroconversion had “definite” acute HCV; b) HCV-seropositive patients with a documented seronegative status within six months and an ALT >7xULN also had “definite” acute HCV; c) those without documentation of prior testing within 6 months who were HCV-seropositive with detectable HCV-RNA and ALT >7xULN had “probable” infection; d) those without documentation of prior testing within 6 months who were HCV-seropositive with detectable HCV-RNA and ALT <7xULN had “possible” infection.
For our novel diagnostic approach, we determined whether serial monitoring of HCV-RNA could better define the probability of acute infection in patients who did not have documentation of prior serologic testing. In this “dynamic” model, patients with recent onset of high-risk behaviors were categorized in terms of probability of acute HCV infection as follows: a) patients who had spontaneous clearance were classified as having “definite” acute HCV infection b) patients with HCV-RNA fluctuations >1 log were defined as “high probability”; c) patients with HCV-RNA fluctuations <1 log were classified as “moderate” or “low probability” based on whether their peak ALT was greater or less than 7xULN; d) patients with any single HCV-RNA level <105 IU/ml were defined as having “high probability” of acute infection.
We conducted retrospective chart reviews of incarcerated patients who were evaluated for chronic HCV infection during this same study period to determine the viral “set point” in a parallel cohort. The same HCV testing methodology used in the acutely-infected patients was applied. Exclusion criteria included HIV infection.
We retrospectively assessed the prevalence of low-level viremia and viral load fluctuations within a separate cohort of community/incarcerated patients diagnosed with “definite” acute HCV presenting to Lemuel Shattuck Hospital (LSH) or Massachusetts General Hospital, partially reported in [29, 34] (Table 1).
All study subjects gave written informed consent. Consent was waived for the de-identified chart review of viral loads from chronically-infected individuals. These protocols conform to 1975 Helsinki guidelines for the conduct of human research and were approved by each hospital’s Institutional Review Board (IRB). The LSH IRB includes a prisoner advocate. Patient consent forms included contact information for a second prisoner advocate (not involved in this research) to address any concerns of the participants.
Comparison of binary features between groups was performed via two-tailed Chi-squared testing or Fisher’s exact test. Comparison of continuous variables between groups was performed via a Mann-Whitney test or a student’s t test. For viral loads between chronically and acutely infected individuals, mean values were compared as more than one measurement from certain individuals during the study. The sensitivity, specificity, and likelihood ratio of proposed diagnostic criteria was performed via nonparametric methods. Figures and analyses were performed by GraphPad Prism-4 (GraphPad Software, La Jolla, CA).
Among the 3,248 inmates who were screened shortly after incarceration, 54 were enrolled as “high-risk”; 37 patients were diagnosed with acute HCV infection. All patients were HIV-seronegative. Sixteen patients had symptoms of hepatitis (43%), including RUQ pain (n=13), fatigue (n=12), nausea (n=8), vomiting (n=4) and jaundice (n=2). Other causes of acute hepatitis, including HAV and HBV were ruled out; alcoholic or cocaine-related hepatitis were unlikely due to the extended period between incarceration and study enrollment (mean= 42 days, IQR [21–51]). Three patients reported a history of heavy alcohol use prior to incarceration.
All study patients had evidence of HCV antibody on enrollment and reported IDU as a risk factor. The majority of patients (73%) never had prior serologic testing. Of the patients with self-reported previous testing, only three could be confirmed (subjects 8–10). These three were the only study participants who were classified as having “definite” acute HCV based on standard definitions (see Methods). Two of these three patients had significant aminotransferase abnormalities, but no detectable HCV-RNA at presentation, consistent with recent resolution of viremia. Nineteen other patients were classified as “probable” and 15 others as “possible” cases based on CDC guidelines.
We subsequently determined whether virologic parameters, such as low-level viremia and/or viral load fluctuations may be prevalent in acute infection. We first tested this concept by retrospectively determining the prevalence of these virologic parameters in a separate cohort of 21 individuals with “definite” acute HCV infection (10 of 21 previously reported ). All subjects had documented seroconversion within 6 months of presentation or achieved spontaneous clearance of viremia; all were HIV-seronegative. Within this well-defined cohort, low-level viremia < 105 IU/mL was present in 17 of 21 patients (81%) and viral load fluctuations were noted in 18 of 21 (86%) (Table 1).
Next, we prospectively determined the prevalence of these virologic parameters in the 35 HCV RNA seropositive incarcerated patients who did not have documented serologic testing within six months of presentation. Consistent with the findings from the acute seroconverter cohort, these virologic parameters were also common in our incarcerated cohort. Specifically, 27 case patients (77%) had documentation of at least one HCV-RNA level <105 IU/mL (mean 17,581 IU/ml; range 682–86,315 IU/ml) (Table 2). In contrast, only 13 percent of 623 control patients with chronic HCV infection, had an HCV-RNA level <105 IU/mL (Figure 1). Using this cutoff, a positive HCV viral load with a value < 105/IU/mL showed a sensitivity of 60% and a specificity of 87% for acute/early HCV, with a likelihood ratio of 4.7 (ROC AUC= 0.84). Furthermore, of the 25 patients who underwent serial HCV-RNA testing, the prevalence of viral load fluctuations (>1 log) was 36 percent (Table 2 and Figure 2). The latter was found to be less frequent in the incarcerated cohort than in the acute HCV seroconverter group (Chi-squared, p=0.002) (Table 3). Whether this may be related to the timing of their diagnosis relative to their onset of infection is unclear. Patients within the incarcerated cohort presented later in the course of infection (median 5.0 months, IQR [2–7]) when compared to those with documented seroconversion, where the median estimated duration of infection was only 3.0 months, IQR [1–5] (Table 3). They also had a lower rate of symptomatic disease, which may have resulted from our systematic risk-based approach (43% vs. 86%, p<0.05).
Using the criterion of low-level viremia, 22 of 35 patients (63 percent) had an HCV-RNA <105 on their initial screening. The utility of serial monitoring as a guide to diagnosis was most apparent in the three patients with “possible” acute infection per CDC criteria (patients 1, 5, 13) who presented with normal aminotransferases at baseline. Two of these patients had low levels of viremia (<105) with subsequent virologic clearance; one was lost to follow-up (Figure 3).
We also carefully examined the subset of patients with persistent viremia who did not meet CDC guidelines for ALT abnormality (e.g., ALT < 7x ULN). Of these 11 patients, the median HCV RNA was 44,025 IU/ml (range 682–810,000 IU/ml; 9 patients had evidence of HCV-RNA levels <105). One of these patients also had evidence of viral load fluctuations of 1.6 log.
In summary, with serial monitoring, we were able to re-classify the vast majority of the patients with “possible” or “probable” acute HCV as “definite” (n=8); “high probability” (n=20); “moderate probability” (n=5) and “low probability” (n=2) using these novel virologic parameters (Table 2). The veracity of these findings is reflected by the high SVR rates achieved in those who accepted therapy .
We have adopted a novel approach to the diagnosis of acute HCV infection, which exploits two virologic parameters that are distinctly uncommon among patients with longstanding disease. Several studies in chronically-infected patients have demonstrated that HCV-RNA levels are relatively stable over long periods of follow-up, varying by less than 0.5 log in the majority of infected persons [17–20]. In contrast, studies of acute HCV pathogenesis have highlighted marked virologic fluctuations in early infection [22, 23, 24], as were seen in over one-third of our patients. Interestingly, we observed that subjects identified in the seroconverter cohort were more likely to demonstrate viral load fluctuations than their prison counterparts, possibly related to their earlier identification within their infection. Variations in viral load are likely related to immunologic control exerted by T-cells in early virus-host interactions [7, 39].
Secondly, we demonstrate that low levels of viremia were also frequently observed in our study participants; in contrast, low HCV-RNA levels were seen in a minority of our controls with chronic infection, consistent with other reports . Low-levels of HCV-RNA may reflect transient immunologic control, early ramp-up of viremia, or viral decline preceding spontaneous clearance, depending on when the patient is evaluated within their disease course. Over time, studies have shown that HCV-RNA will tend to remain stable or increase as infected individuals enter the chronic phase .
Although low levels of HCV-RNA and viral load fluctuations have been demonstrated in multiple reports of patients with acute HCV infection, these laboratory parameters have not been integrated into diagnostic criteria [14, 23, 27–29]. In our well-defined cohort of “definite” seroconverters, we have demonstrated the validity of applying these virologic parameters as supportive laboratory criteria. We propose that standard diagnostic criteria for acute HCV infection should be expanded to include these virologic parameters, which perform at least as well as ALT elevations, and may be even more specific. We view the diagnosis of acute HCV infection as a dynamic process, best determined by serial monitoring of HCV-RNA levels over a 10 to 12-week period. This approach was particularly useful in patients without prior HCV antibody testing. Using a threshold of ALT >7x ULN as major criteria for the diagnosis of acute HCV infection would have missed 15 patients in the prison cohort (including four who spontaneously cleared), showing the insensitivity of this criterion. In contrast, of patients with persistent viremia and ALT <7x ULN, the majority exhibited either viral load fluctuations or low levels of viremia.
The vast majority of IDUs with suspected acute hepatitis C can be classified with a high level of confidence based on this straightforward approach, which incorporates ascertainment of risk and serial laboratory monitoring (Figure 4). Thirteen of 15 patients with “possible” acute HCV infection on presentation were reclassified as “definite” or “high probability” based on either virologic clearance, viral load fluctuations >1 log, or low levels of viremia (<105). The validity of this approach is supported by three lines of evidence: a) the high prevalence of these virologic parameters in our cohort of documented seroconverters and in other published studies; b) the low prevalence of these parameters in prior studies of chronically-infected individuals and in our control group of chronic HCV patients; and c) preliminary data demonstrating high SVR in the inmates who have completed HCV therapy for persistent viremia . The application of these proposed criteria would have the greatest impact on patients with persistent viremia, who should be promptly referred for treatment. These criteria may also help improve acute HCV surveillance programs on a local/national level.
We propose the creation of an expanded diagnostic scoring system, using weighted points for the various factors we have considered in our model. This classification system would be based on virologic outcomes, changes in viremia (HCV-RNA fluctuations < or >1 log), levels of viremia (< or >105 IU/ml), and aminotransferase testing. This system would particularly enhance the diagnosis of HCV-seropositive patients without prior testing and would facilitate timely therapeutic interventions. However, this approach needs to be validated in larger cohorts.
Our data are limited by a lag in initial laboratory testing in some participants, such that their initial viremia was not captured due to spontaneous clearance prior to testing. Furthermore, we may have overestimated the proportion of patients who spontaneously cleared due to the lower limits of HCV-RNA detection by our assays (<615 IU/ml by Versant and <600 IU/ml Roche Amplicor). In addition, we suspect our case finding was not optimal due to lack of universal screening of inmates entering the two intake sites. Despite this limitation, we identified 37 patients with acute hepatitis C infection within 15 months of screening through risk assessment.
In conclusion, in IDUs with recent onset of risk-taking behaviors, we suggest serial HCV-RNA monitoring to assess viral load fluctuations and height of viremia as useful parameters in the diagnosis of acute HCV infection. This approach is particularly helpful in those patients who do not have baseline HCV antibody testing, as is common in IDUs.
We especially thank the individuals who consented to take part in this study. We acknowledge Warren Ferguson and Arthur Brewer of the University of Massachusetts Correctional Health and the providers at MCI-Framingham and MCI-Concord for their support.
Funding sources: National Institutes of Health / National Institute of Allergy and Infectious Diseases (Hepatitis C Cooperative Center U19 AI066345, K23 AI054379 to AYK, Harvard University Center for AIDS Research P30 AI060354).
Conflict of Interest Statement: AYK, CEB, EHN, LLR, and MJB report no conflicts of interest. BHM is on the speaker’s bureau of Roche Pharmaceuticals and advisory boards of Merck and Vertex Pharmaceuticals. RTC receives research grant support from Roche Labs and Schering-Plough.
Article Summary: The diagnosis of acute infection in the HCV-seropositive patient is strengthened by the use of virologic parameters that are uncommon in chronic disease. Viral load fluctuations and low levels of HCV-RNA should be incorporated into standard diagnostic criteria.