As a result of shared routes of infection, co-infection with HIV and hepatitis C virus (HCV) is common . HIV and HCV are each associated with higher than expected rates of comorbid conditions such as anemia , alcohol and drug disorders [3,4], and depression [5–7]. Other comorbidities (e.g. hypertension), chronic obstructive pulmonary disease, and coronary artery disease (CAD) may also impact upon HCV treatment eligibility and outcomes, and could increase in prevalence as HIV patients survive longer and continue to age.
Many comorbid conditions are contraindications to HCV treatment . Relative contraindications include CAD, uncontrolled diabetes, severe psychiatric disorder, pancreatitis, a history of poor medication adherence, and active or past intravenous drug use or alcohol disorder. Absolute contraindications include hypersensitivity to interferon or ribavirin, autoimmune disease, hemoglobinopathies, active opportunistic infections, and decompensated liver disease . In addition, comorbid conditions may complicate treatment indirectly by increasing the potential for adverse medication side-effects. Consequently, disproportionately more patients with comorbid conditions may remain untreated despite meeting criteria for the treatment of HCV.
Treatment of the co-infected patient, especially those with comorbidity, must be individualized because of the potential for complications [10,11], including the exacerbation of depressive symptoms [12,13], anxiety, and irritability [14,15]. Ribavirin therapy can produce dose-limiting anemia. Patients with pre-existing anemia may be unable to tolerate ribavirin, even if the dose is reduced. Symptoms of CAD may be exacerbated by ribavirin-induced anemia. Active alcohol and drug use disorders may increase HCV viremia, increase the risk of hepatic fibrosis, and decrease treatment response [16–19].
Despite the high degree of comorbidity observed in both HIV and HCV-infected individuals, and its potential impact on HCV treatment eligibility and outcomes, the prevalence of comorbidity among HIV and HCV-co-infected individuals has not been well described. In order to estimate the prevalence of co-infection and to determine whether comorbid conditions are more common among co-infected compared with HIV-mono-infected patients, this study utilized information from the largest single provider of HIV care in the United States: the Veterans Health Administration. The Veterans Health Administration, which provided care to over 19 000 HIV patients in fiscal year 2002, has one of the foremost electronic health information systems in the country, and is a unique source of data on which to conduct this study .
We identified all HIV-positive veterans in care using the methodology of Fasciano et al. . We searched the VA electronic data files for all patients who had received a documented HIV diagnosis between October 1997 and September 2002. Information on HCV and comorbid conditions was also extracted. Both inpatient and outpatient treatment files were utilized. The Institutional Review Boards of Yale University and the VA Connecticut Healthcare System approved the project. All veterans with HIV being treated between 1997 and 2002 were included in the analysis.
We searched the files for records containing International Classification of Diseases 9th Edition-Clinical Modifications (ICD-9-CM) HIV codes 042–044, 044.9, 795.8, or V08 (asymptomatic HIV) and HIV-related diagnosis-related group codes 488–490. Patients with records meeting any of these inclusion criteria were included in the sample. ICD-9 and diagnosis-related group codes for hepatitis C and comorbid medical and psychiatric conditions were extracted in a similar manner . We identified comorbidity diagnoses present in the 5 years before the index visit. The classification of comorbid alcohol, drug, and psychiatric conditions from ICD-9 codes in the electronic data was based on the methods of Druss and Rosenheck .
We assessed the ability of the VA electronic data to identify veterans receiving HIV care correctly. Utilizing a sample of 1851 veterans, 1064 in care for HIV and 787 in care for other medical conditions, all of whom are prospectively enrolled in a study being conducted by the authors, we found that the electronic data correctly identified 100% of the veterans in HIV care (i.e. sensitivity), the specificity was 96.1%, the positive predictive value was 97.2%, and the negative predictive value was 100%.
We also assessed the validity of VA electronic data in identifying HCV and other conditions among a sample of 881 veterans with HIV. Using formal chart extraction as the gold standard for all conditions save anemia (hemoglobin was the gold standard) and HCV (HCV antibody positive was the gold standard for this analysis), the electronic data demonstrated a fair degree of accuracy in identifying patients being treated for each of the conditions assessed, and overall good sensitivity (see Appendix). Kappa statistics, a measure of chance-corrected agreement , revealed that electronic and chart review as well as laboratory data were in moderate agreement .
We calculated prevalence rates based upon the observed proportion of patients, overall and by co-infection status, for each comorbid condition of interest. Logistic regression was used to calculate age and race-adjusted odds ratios (AOR) for each comorbid condition, comparing co-infected with HIV-mono-infected patients. To illustrate more clearly the association between comorbidity and co-infection, a factor analysis was conducted on the comorbid conditions assessed in this study. The association between comorbid conditions was calculated using tetrachoric correlation coefficients, a measure of association between dichotomous variables . The tetrachoric correlation matrix was analysed to extract factors that explained the maximum amount of variance of the comorbid conditions. Only factors with eigenvalues greater than 1.0 were retained for analysis. Factors scores were subjected to Varimax rotation in order to facilitate interpretation. In this study, variables with factor loadings greater than 0.40 were determined to be loaded on a factor. Complex variables, those with high loadings on more than one factor, were retained in the final solution in order to portray the factor structure more accurately. The rotated factor scores obtained were dichotomized so that veterans with scores in the highest quintile were considered to have the syndrome that the factor represented. Statistical significance for all tests was defined as a two-tailed P value less than 0.05. All statistical analyses were performed using SAS software (version 8; The SAS Institute, Cary, NC, USA).
The sample included 25 116 HIV-infected veterans. The median age was 47 years (range 18–99). The veterans in this sample were primarily non-white (31% African-American and 38% Hispanic), and 97% were male. Eighteen per cent of the sample (n = 4489) had evidence of HCV co-infection. Co-infected veterans were significantly more likely to be male (98 versus 94%, P < 0.001), and to be non-white (69 versus 61%, P < 0.001; Table 1).
HCV-co-infected veterans had a higher unadjusted prevalence of most of the comorbid conditions assessed (Table 2). The AOR for comorbid conditions among co-infected patients compared with mono-infected patients ranged from 1.01 for CAD to 5.05 for drug disorders. Co-infected veterans had a significantly higher AOR for alcohol disorders, drug disorders, depressive disorders, posttraumatic stress disorder, schizophrenia, bipolar disorder, as well as pancreatitis, anemia, diabetes, stroke/transient ischemic attack, chronic obstructive pulmonary disease/asthma, hypertension, and congestive heart failure.
Overall, 67% of veterans in this sample had evidence of one or more comorbid conditions. HCV-co-infected veterans were significantly more likely to have one or more comorbid conditions compared with HIV-mono-infected veterans (85 versus 63%, P < 0.0001). Co-infected veterans also had a significantly higher median number of comorbid conditions (two versus one, P < 0.0001). Nearly one-half of HCV-co-infected veterans had three or more comorbid conditions, whereas less than one-quarter of HIV-mono-infected veterans did (Fig. 1).
A factor analysis on the comorbid conditions assessed in this study yielded three factors explaining 62% of the total variance of the conditions. The factor accounting for the highest proportion of the variance (28%) had strong loadings by alcohol, drug, and the four psychiatric disorders, and was therefore termed the ‘mental disorder’ factor. The second factor, with loadings from CAD, congestive heart failure, and hypertension, accounted for 25% of the variance, and was regarded as a ‘medical disorder’ factor. The third, with high loadings by pancreatitis, anemia, and alcohol disorders, explained 9% of the variance, and was deemed the ‘alcohol-related complications’ factor (Table 3). Alcohol disorder was the only variable that loaded on two factors, with a higher loading (0.70) on the ‘mental disorder’ than the ‘alcohol-related complications’ factor (0.51). Such dual loadings indicate that alcohol disorders is related to two distinct clusters of conditions, and that each cluster is in turn unrelated to the other.
HCV-co-infected veterans were significantly more likely to have scores in the upper quintile on one or more factors compared with HIV-mono-infected veterans (60 versus 43%, P < 0.0001). Co-infected veterans were significantly more likely to have high scores on the mental disorders and the alcohol-related complications factors (33 versus 17%, and 32 versus 18%, respectively; P < 0.0001 in both cases), but were significantly less likely to have high scores on the medical disorders factor (14 versus 21%, P < 0.0001). Logistic regression models adjusting for the effects of age and race revealed that co-infected veterans were significantly more likely to have the mental disorders factor (AOR 2.78, P < 0.0001), and the alcohol-related complications factor (AOR 2.12, P < 0.0001), but were significantly less likely to have the medical disorders factor (AOR 0.88, P < 0.001). Results were comparable in analyses controlling for the presence of the other factors (e.g. alcohol-related factor, controlling for mental and medical disorders factors).
There was a substantial degree of overlap between the three factors (Fig. 2). Among all HIV-mono-infected veterans (n = 20 627), 8% had high scores on the mental disorders factor without high scores on the medical or alcohol-related conditions factors; 12% had medical only; and 12% had high scores on the alcohol-related conditions factor only (Fig. 2a). One per cent (center of Fig. 2a) of mono-infected veterans had a combination of mental disorders, medical disorders, and the alcohol-related conditions factors. Among all HIV/HCV-co-infected veterans (n = 4489), the corresponding proportions were: 19% for mental disorders factor alone; 6% for medical disorders; and 18% for alcohol-related conditions only. Three per cent of co-infected veterans had mental disorders, medical disorders, and alcohol-related conditions factors combined (Fig. 2b). Veterans with co-infection were significantly more likely to have high loadings on all three factors combined (AOR 2.65, P < 0.001).
The prevalence of HCV among veterans in care for HIV was nearly sevenfold higher than that reported among adults in the general US population (18 versus 2.7%), and more than triple that of veterans in general (5.4%) [27,28]. HCV-co-infected veterans were significantly older and more likely to be of minority race compared with HIV-mono-infected veterans, both characteristics of a large proportion of veterans served by the VA that are also associated with a higher prevalence of a variety of comorbid conditions. Combined, this suggests a potential for a substantial increase in future morbidity and mortality from HCV-related disease among HIV-infected veterans.
HCV-co-infected veterans were more likely to have comorbid conditions, and had a higher likelihood of having multiple, overlapping comorbidities. The high prevalence of multiple comorbidities among co-infected veterans, especially alcohol, drug, and psychiatric disorders, may limit treatment options for many co-infected patients, based upon current HCV treatment guidelines . In addition, co-infected patients had a higher prevalence of comorbid conditions not considered direct contraindications to treatment (e.g. hypertension). However, these conditions may impact HCV treatment options by increasing the amount of medications a patient requires, thereby increasing the risk of adverse effects and non-adherence.
In the factor analysis, alcohol disorder loaded on two factors: mental disorders and alcohol-related complications. The implication of the dual loading of alcohol disorder is that the ill effects of hazardous alcohol use impacts multiple domains of illness. This is not unexpected. Alcohol disorders commonly co-occur with mental disorders, and may represent a common predisposition or the use of alcohol to treat psychiatric symptoms. Conversely, pancreatitis and anemia are direct biological complications of hazardous levels of alcohol use. An important implication of this finding is that treatment of alcohol disorders has the potential to impact multiple levels of comorbidity.
This study has several strengths, including a large racially and geographically diverse sample. The sample also includes a large number of older patients, who already comprise a substantial and increasing proportion of HIV patients . In addition, this sample includes many patients likely to have been exposed to both HIV and HCV through injection drug use. This is especially relevant given the high prevalence of HCV co-infection among injection drug users, and the central role that injection drug use has played in the spread of both HIV and HCV.
Additional strengths of this study are our examination of the validity of the ICD-9 codes used to identify veterans with HIV, HCV, as well as the comorbid conditions that may be contraindications for HCV treatment, and the use of factor analysis to reduce the set of intercorrelated comorbid conditions into a smaller number of factors that can be used more readily to examine the complex association between HCV co-infection and comorbidity. We demonstrated that VA electronic data corresponds moderately well to chart review and laboratory test results for a variety of conditions, and that it could be used to identify patients with these conditions. We also demonstrated that factor analysis concisely captures a substantial amount of the variance among the many comorbid conditions, and logically represents those conditions using three factors. Veterans with HCV co-infection were more than twice as likely to have both the mental disorders and the alcohol-related conditions factors, even after adjusting for age and race. This finding succinctly conveys the domains of comorbidity that may adversely affect treatment for HIV and HCV-co-infected patients.
There are several limitations to this study. First, we were unable to conduct formal patient evaluations, and instead relied upon diagnoses in the patients’ electronic records. This may result in an underestimation of the actual prevalence in this sample if some diagnoses were not recorded. The use of electronic data may also result in an overestimation if patients with co-infection were more likely to have been screened and treated for comorbid conditions, or if individuals with comorbid conditions are more likely to be screened for HCV. Although not all-inclusive, the comorbid conditions we examined represent high impact and high prevalence conditions likely to be experienced by a significant number of mono- and co-infected patients, especially as these patients age over time. Because of the lack of laboratory and pharmacy information, we were unable to adjust for differences in HIV disease status and treatment. Both HIV disease status and treatment could result in many of the comorbid conditions. Any analysis examining treatment in the setting of comorbidities will need to take into account the limitations on treatment the comorbidities themselves may create.
ICD-9 codes were fairly insensitive for the diagnosis of hepatitis C in this study. In a sub-study analysis, we compared hepatitis C antibody-positive patients with and without ICD-9 codes and found no significant difference in the prevalence of comorbidities. This suggests that our results would be strengthened by the identification of patients with HCV antibodies but without HCV ICD-9 codes.
Our findings indicate that a higher proportion of HIV and HCV-co-infected patients are affected by comorbid conditions than HIV-mono-infected patients, and that multiple comorbidities are exceedingly common among co-infected patients. Providers are frequently faced with the dilemma of when to treat the comorbid condition relative to treatment for the HIV and HCV infections. Clinical management is further complicated by the fact that treatment for the comorbidities may impact the treatment of HIV and HCV because of the increased risk of adverse side-effects.
Efforts must be made to increase the availability and tolerability of treatment for co-infected patients through the development of treatment strategies that are not adversely impacted by these comorbidities. Further research is needed to determine the extent to which treatment of comorbidities increases the rate of successful HCV treatment, and whether the treatment of comorbidities improves HCV-related outcomes. Such research will also need to address questions regarding the order in which multiple comorbidities are best treated, and the impact of comorbidities not directly considered contraindications to treatment.
Sponsorship: This study was funded by the National Institute on Alcohol and Alcohol Abuse (3U01 AA 13566), National Institute of Aging (K23 AG00826), Robert Wood Johnson Generalist Faculty Scholar Award, an Inter-agency Agreement between National Institute of Aging, National Institute of Mental Health, unrestricted educational grants from GlaxoSmithKline, Inc. and Agouron (A.C.J.).
1. Soriano V, Rodriguez-Rosado R, Garcia-Samaniego J. Management of chronic hepatitis C in HIV-infected patients. AIDS 1999; 13:539–546.
2. Mildvan D. Implications of anemia in human immunodeficiency virus, cancer, and hepatitis C virus. Clin Infect Dis 2003; 37(Suppl. 4):293–296.
3. Conigliaro J, Gordon AJ, McGinnis KA, Rabeneck L, Justice AC, for the Veterans Aging Cohort 3-Site Study. How harmful is hazardous alcohol use and abuse in HIV infection: do health care providers know who is at risk? J Acquir Immune Defic Syndr 2003; 33:521–525.
4. Galvan FH, Bing EG, Fleishman JA, London AS, Caetano R, Burnam MA, et al
. The prevalence of alcohol consumption and heavy drinking among people with HIV in the United States: results from the HIV Cost and Services Utilization Study. J Stud Alcohol 2002; 63:179–186.
5. Cruess DG, Evans DL, Repetto MJ, Gettes D, Douglas SD, Petitto JM. Prevalence, diagnosis, and pharmacological treatment of mood disorders in HIV disease. Biol Psychiatry 2003; 54:307–316.
6. Kilbourne AM, Justice AC, Rollman BL, McGinnis KA, Rabeneck L, Weissman S, et al
. Clinical importance of HIV and depressive symptoms among veterans with HIV infection. J Gen Intern Med 2002; 17:512–520.
7. Dieperink E, Willenbring M, Ho SB. Neuropsychiatric symptoms associated with hepatitis C and interferon alpha: a review. Am J Psychiatry 2000; 157:867–876.
8. Fleming CA, Craven DE, Thornton D, Tumilty S, Nunes D. Hepatitis C virus and human immunodeficiency virus coinfection in an urban population: low eligibility for interferon treatment. Clin Infect Dis 2003; 36:97–100.
10. US Public Health Service and Infectious Disease Society of America (IDSA). 2001 USPHS/IDSA guidelines for the management of opportunistic infections in persons infected with HIV
. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service; 2001 Nov 28.
11. Soriano V, Sulkowski M, Bergin C, Hatzakis A, Cacoub P, Katlama C, et al
. Care of patients with chronic hepatitis C and HIV co-infection: recommendations from the HIV-HCV International Panel. AIDS 2002; 16:813–828.
12. Gallucci G, Smolinski J. Treatment contracts for patients with hepatitis C, psychiatric illness, and substance abuse. Psychosomatics 2001; 42:353–355.
13. Dobmeier M, Frick E, Frank S, Franke C, Wolfersdorf M. Schizophrenic psychosis: a contraindication for treatment of hepatitis C with interferon alpha? Pharmacopsychiatry 2000; 33:72–74.
14. Trask PC, Esper P, Riba M, Redman B. Psychiatric side effects of interferon therapy: prevalence, proposed mechanisms, and future directions. J Clin Oncol 2000; 18:2316–2326.
15. Schafer M, Boetsch T, Laakmann G. Psychosis in a methadone substituted patient during interferon-alpha treatment of hepatitis C. Addiction 2000; 95:1101–1104.
16. Pessione F, Degos F, Marcellin P, Duchatelle V, Njapoum C, Martinot-Peignoux M, et al
. Effect of alcohol consumption on serum hepatitis C virus RNA and histological lesions in chronic hepatitis C. Hepatology 1998; 27:1717–1722.
17. Strathdee SA, Palepu A, Cornelisse P, Yip B, O'Shaughnessy MV, Montaner JS, et al
. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998; 280:547–549.
18. Kresina TF, Flexner CW, Sinclair J, Correia MA, Stapleton JT, Adeniyi-Jones S, et al
. Alcohol use and HIV pharmacotherapy. AIDS Res Hum Retroviruses 2002; 18:757–770.
19. Schiff ER. The alcoholic patient with hepatitis C virus infection. Am J Med 1999; 107(Suppl. 6B):95s–99s.
20. Corrigan JM, Eden J, Smith BM, editors. Leadership by example: coordinating government roles in improving health care quality
. Committee on Enhancing Federal Healthcare Quality Programs. Washington, DC: The National Academies Press; 2003.
21. Fasciano N, Cherlow A, Turner B, Thornton C. Profile of Medicare beneficiaries with AIDS: application of an AIDS case finding algorithm. Health Care Financ Rev 1998; 19:19–38.
22. McGinnis KA, Fine MJ, Sharma RK, Skanderson M, Wagner JH, Rodriguez-Barradas MC, et al
. Understanding racial disparities in HIV using data from the Veterans Aging Cohort 3-Site Study and VA administrative data. Am J Public Health 2003; 93:1728–1733.
23. Druss BG, Rosenheck RA. Locus of mental health treatment in an integrated service system. Psychiatr Serv 2000; 51:890–892.
24. Fleiss JL. Measuring nominal scale agreement among many raters. Psychol Bull 1971; 76:378–382.
25. Landis RJ, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174.
26. Drasgow F. Polychoric and polyserial correlations.
In Kotz L, Johnson NL, editors. Encyclopedia of statistical sciences
, Vol. 7. New York: Wiley; 1988. pp. 69–74.
27. Alter MJ, Kruszon-Moran D, Nainan OV, McQuillan GM, Gao F, Moyer LA, et al
. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med 1999; 341:556–562.
28. Roselle GA, Danko LH, Kralovic SM, Simbartl LA, Kizer KW. National hepatitis C surveillance day in the Veterans Health Administration of the Department of Veterans Affairs. Military Med 2002; 167:756–759.
29. Justice AC, Landefeld SC, Asch SM, Gifford A, Whalen C, Covinsky K. Justification for a new cohort study of people aging with and without HIV infection. J Clin Epidemiol 2001; 54(Suppl. 1):S3–S8.
Prevalence of comorbid disorders by source of information, percentage agreement, and kappa statistics for agreement beyond chance, among 881 veterans in the validation sample.
© 2005 Lippincott Williams & Wilkins, Inc.