Symptom and Comorbidity Burden in Chronic Disease: Comparison of HIV- Infection and Diabetes Mellitus in Aging Patients

Eva Wolf1, Christian Hoffmann2, Knud Schewe2,12, Stephan Klauke3,12, Robert Baumann4, Martin Karwat5, Frank Schlote6, Franz Mosthaf7, Hans Heiken8, Axel Baumgarten9, Albrecht Ulmer10, and Hans Jaeger11,12 for the dagnae 50/2010 study group 1MUC Research, Munich, Germany 2ICH Study Center, Hamburg, Germany 3Infektiologikum Frankfurt, Frankfurt, Germany 4Private Practice Dr. R. Baumann, Neuss, Germany 5Private Practice Dr. M. Karwat, Munich, Germany 6Praxisgemeinschaft Turmstrasse, Berlin, Germany 7Private Practice for Hematology, Oncology and Infectious Diseases, Karlsruhe 8Praxis Georgstrasse, Hannover, Germany 9Medical Center for Infectious Diseases (MIB), Berlin, Germany 10Private Practice Dres. Ulmer/Frietsch/Mueller/Roll, Stuttgart, Germany 11MVZ Karlsplatz – HIV Research and Clinical Care Centre, Munich, Germany 12dagnae e.V., The German Association of Physicians Specialized in HIV Care, Berlin, Germany


Introduction
The number of people living with HIV/AIDS in the 50-year-plus age group is rising, especially in industrialized nations. With the widespread use of highly active antiretroviral therapy (HAART), physicians and patients are facing new challenges in the treatment and management of age-related comorbidities [1][2][3][4][5]. Several cohort studies showed that HIV-infected patients are at increased risk for individual comorbidities, i.e. for bone [1,[3][4][5], renal [1,6,7], neurocognitive and cardiovascular disorders [1,8,9] as well as non-AIDS-defining malignancies [10]. The relative contribution of HIV-infection and HIV-related factors such as inflammation, immune activation and altered coagulation to the development of different comorbidities in aging individuals is still inconclusive. Other risk factors, life style factors or the use of antiretroviral drugs may contribute to excess morbidity [2,8,[11][12][13][14][15][16][17].
With this study, we tried to identify the differences in the total burden of disease between HIV-infected patients in comparison to two HIV-negative comparator groups. The first group included patients without any severe chronic or malignant disease, the second included patients with diabetes mellitus type 2, another chronic disease representing a major global cause of premature morbidity and excess mortality [18][19][20].

Study design and study population
50/2010 (the 50-year-plus age group in 2010) was a noninterventional, multi-center nationwide prospective 18-months cohort study initiated by The German Association of Physicians Specialized in HIV Care (dagnae e.V.) comparing HIV-infected patients with two HIV-negative patient groups, i.e. patients with diabetes mellitus type 2 (DM) and control patients. Inclusion criteria were an age ≥50 years and the absence of any life-threatening or malignant disease requiring treatment. Additional inclusion criteria for HIV-infected patients were chronic HIV-infection and no acute opportunistic infection at the time of study entry.
Inclusion criteria for the DM group was clinical diagnosis of DM verified by its documentation in the patient file or HbA1c levels ≥6.5%, fasting plasma glucose levels ≥126 mg/dL or the use of anti-diabetic drugs.
Inclusion criteria for control patients were the absence of any severe or unstable chronic disease at baseline. Health conditions and comorbidities that did not meet the criteria for exclusion in either study group were hypertension, dyslipidemia, depression, history of heart attack or cerebrovascular event (if more than one year ago) and chronic infection with hepatitis C or hepatitis B virus (if diagnosis was obtained more than one year ago). Concomitant hepatitis C infection requiring treatment was excluded. Patients with a history of intravenous drug use receiving replacement therapy were not excluded.
The recruitment of women was stopped in a group if they exceeded 20% (according to the gender distribution of the HIV-infected population in Germany).

Regulatory requirements
Prior to the conduct of the study, approvals of the institutional review boards, i.e. the competent ethics committees, were obtained for all study sites. All subjects had to give written informed consent before inclusion in the study.

Data collection
Biological, clinical and psychosocial parameters were collected at baseline, months 6, 12 and 18. The case report form included sociodemographic variables, HIV-related history, vital signs, body weight, laboratory parameters, relevant comorbidities and conditions, and concomitant medication. At each visit patients had to complete questionnaires including life style factors and symptoms of aging. For the male study population, the Aging Males' Symptoms (AMS) scale was used, a self-reported, 17-item questionnaire including a somatic, a psychological and a sexual subscale. The total AMS score has a range from 17 to 85; a total score ≥37 indicates moderate or severe subjectively perceived complaints [21]. Women were interviewed regarding the onset of menopause and related symptoms.
As an accepted indicator of weakness/frailty and health status, grip strength of the dominant hand was measured using a dynamometer (Saehan Dynamometer, model SH5001) [22,23]. The prevalence of weakness was defined as grip strength of 30 kg or less for males and 18 kg or less for females [24,25]. Factors associated with weak grip strength were evaluated using logistic regression analysis. Covariables considered were HIV-infection, DM, BMI (<20 versus (vs) ≥20 kg/ m 2 ), number of documented comorbidities (≥2 vs <2), age (≥65 vs <65 years), gender and physical activity (<2x vs ≥2x 45 min/week). In sensitivity analyses, hepatitis B virus (HBV) and hepatitis C virus (HCV) infection were included as covariables in the model to adjust for potential confounding.
The Hospital Anxiety and Depression Scale (HADS), a selfscreening questionnaire, was used to assess anxiety and depression.
Health-related quality of life was evaluated using the 36-item Short Form Health Survey questionnaire (SF-36). Results will be presented elsewhere.
The 10-year risk for coronary heart disease (CHD) including myocardial infarction or coronary death was calculated using the Framingham risk equation considering age, the presence of diabetes mellitus, smoking, JNC-V blood pressure categories, and NCEP total cholesterol categories [26]. A 10-year risk ≥20% or a history of CHD were defined as high risk. Multivariate logistic regression analysis adjusting for age categories, diabetes mellitus, a positive family history for CHD and smoking was used to evaluate whether HIV-infection by itself represents an additional risk for CHD. Due to low numbers, patients aged ≥80 years were excluded from this evaluation.

Statistics
Continuous variables were described as medians with interquartile ranges (IQR). The Mann-Whitney U test and the Kruskal Wallis test were used for comparison of continuous variables between two and more groups. The Wilcoxon signed rank test was used to test for significant changes in continuous variables within groups. Fisher´s exact test and the chi 2 test were used for comparison of frequencies between two and more groups. The McNemar test was used to test for significant changes in frequencies within groups. The P-level for significance was P<0.05 (n.s. = not significant).

Study population
The study population consisted of 761 participants including 255 patients with chronic HIV-infection (HIV group), 249 HIV-negative patients with diabetes mellitus type 2 (DM group) and 257 HIV-negative controls, all aged ≥50 years. The study participants were recruited in 37 German clinics between July 2008 and December 2009. The last followup visit was in August 2011.
Socio-demographic characteristics of the study population are shown in Table 1. Twenty-one point two percent of patients were of female gender. Chronic hepatitis B was present in 2.4%, 0.8% and 1.2% in HIV, DM and control patients while chronic hepatitis C was found in 5.9%, 2.0% and 2.7%, respectively. Replacement therapy in patients with a history of i.v. drug use was used in 2.0%, 0.4% and 2.0% of HIV, DM or control patients, respectively.

Physical constitution, activity and weakness
Physical constitution, physical activity and weakness at study entry are shown in Table 2. In the HIV group, significantly more patients had a low BMI <20kg/m 2 in comparison to the HIV-negative groups (10.5% vs ≤1%); this was true for men and females. In contrast, high BMI was most prevalent in the diabetes mellitus group. BMI did not change significantly during follow-up in either group. Physical activity was comparable between the HIV group and the control group (about half of the patients reported physical activity for at least 2 times per week for ≥45 min), but was lower in the DM group.
In male patients, the median grip strength was significantly lower in the HIV group than in the DM group or in the control group. In women, we did not observe significant differences. Grip strength did

Symptoms of aging
The Aging Males' Symptoms (AMS) questionnaire was completed by 514 male patients at baseline and 425 patients at month 12. At baseline, median AMS summary scores were 37.0, 33.3 and 31.0 in HIV, DM and control groups, respectively. There were significant differences between the HIV and the control group (P<0.001) and between the DM and the control group (P=0.0039) but not between the HIV and DM group.
AMS summary scores ≥37 indicating moderate or severe complaints were significantly more prevalent in HIV and DM groups than in the control group. The prevalences were 51% in the HIV group (47% in patients with HIV-RNA <50 copies/mL, 75% in patients with HIV-RNA ≥50 copies/mL, P=0.008), 45% in the DM group compared to 29% in the control group (pairwise comparisons with the control group, P<0.0001 and P=0.002, respectively).
HIV-infection (OR 2.2, 95% CI 1.5-3.5) and diabetes mellitus (OR 1.9, 95% CI 1.2-2.9) were independently associated with a high AMS summary score after adjusting for age categories (60-69, 70-79 vs 50-59 years), high BMI (>28 vs ≤28 kg/m 2 ) and stable partnership. Adjusting for HBV-and HCV-infection did not affect these associations. Stable partnership (OR 0.6, 95% CI 0.4-0.8) was also associated with a lower AMS score. By month 12, there was a small but significant increase in AMS summary score of +2 points in each group.
In women, the onset of menopause was similar across the three patient groups. Median age was 50.

Cardiovascular risk factors and risk for coronary heart disease (CHD)
In total, 566 patients had complete data for calculation of CHD risk at the baseline visit (204 HIV, 196 DM, and 166 control patients). Among the three groups, 39.7 %, 58.7 % and 26.5 % were at high risk for CHD, respectively (Figure 1). History of CHD and low plasma HDL (high density lipoprotein) levels were significantly more frequent in the HIV and DM groups than in the control group. In contrast, HIV and control patients had significantly higher total cholesterol and LDL (high density lipoprotein) levels than DM patients who were significantly more frequently on lipid lowering drugs (Table 3). Logistic regression analysis adjusting for gender, age categories, DM, smoking and positive family history for CHD showed that HIV-infection remained independently associated with high CHD risk (OR 1.6, 95% CI 1.0-2.5, P=0.046, Table 4).

Comorbidities, malignancies and other clinical manifestations
The prevalence of comorbidities or other medical disorders at baseline was highest in the DM group, followed by the HIV group. At least two comorbidities or other medical disorders were documented in 76.3% of the DM group, in 67.8% of the HIV group (70% in patients with HIV-RNA <50 copies/mL, 63% in patients with HIV-RNA ≥50 copies/mL) and in 59.1% of the control group (P<0.001) (76.0%, 67.1% and 57.9%, respectively, when excluding HBV-and HCV-infection). Pairwise comparisons showed significant differences between groups. For certain comorbidities, we observed significantly different prevalences in the three groups, with post-hoc tests showing significantly higher prevalences in HIV-infected patients than in control patients, The difference between HIV-infected patients and control patients was significant (P<0.001), as was the difference between DM and control patients (P=0.015). The corresponding relative incidence rates were significantly different from 1. The lower 95% confidence interval limits were 2.1 and 1.2, respectively.

Discussion
This large cohort study is the first to our knowledge to evaluate health conditions and the disease burden in aging HIV-infected patients in comparison to patients with another age-affected condition, diabetes mellitus type 2 and to individuals without any severe chronic disease. HIV-infection was associated with a higher burden of comorbidity, i.e. cardiovascular diseases and malignancies, when compared to patients without severe chronic disease. However, the overall prevalence of all comorbidities and other medical disorders was highest in patients with DM. For certain comorbidities including renal disorders, neurological disorders or history of myocardial infarction, HIV-infected patients were comparably affected. The calculated risk of coronary heart disease was significantly higher in HIV-infected patients (of whom 96% received antiretroviral treatment) than in HIV-negative controls without chronic disease. However, the excess risk attributed to HIVinfection was lower than that attributed to diabetes mellitus type 2. Our results are in concordance with a systematic review including a metaanalysis and a large cohort study showing that people living with HIV are at increased risk of cardiovascular disease beyond that explained by cardiovascular risk factors [27,28].
The frequency and the burden of symptoms as assessed by the Aging Males' Symptoms (AMS) scale were comparable between HIV and DM patients. Both groups had significantly higher AMS summary scores than the control group. High symptoms scores were even more common in HIV-infected patients with detectable plasma viremia than in patients with viral suppression to <50 HIV-RNA copies/ mL. Symptoms of aging included psychological, somatic and sexual complaints.
Weakness, defined as low grip strength, was more prevalent in HIV-infected individuals than in the two HIV-negative groups. Grip strength is widely used as an indicator of physical functioning and is considered as an independent predictor of morbidity and mortality [23,29,30]. After adjusting for age, DM, BMI, and physical activity, the association between weakness and HIV-infection remained significant.
The use of concomitant drugs for the treatment of disorders which are often associated with DM was much higher in the DM group than in the two other groups, i.e. the use of antihypertensive agents, lipid lowering drugs and anticoagulants. Interestingly, despite a higher cardiovascular risk in HIV-infected patients in comparison to control patients, the use of antihypertensive drugs was less common in HIV-   infected patients, and the use of lipid lowering drugs was comparable between HIV-infected and control patients. The use of psychotropic drugs was higher in HIV-infected patients than in the HIV-negative groups.
Although the incidence of malignancies was small during the 18-month follow-up, we observed a significantly higher incidence of non-AIDS-defining malignancies (NADM) in the HIV-infected group than in the control group. These findings are in concordance with data showing a higher incidence of NADM in HIV-infected persons in the HAART era than in the general population, adjusted for age, race, and gender [31,32]. Based on nationwide cohort data, the incidence of cancer among HIV-infected patients in France was 1.4 per 100 person years for the year 2006, with NADM representing 68% of cases and including a large variety of entities. The relative risk for HIV-infected patients when compared to the general population in France was estimated to be 3.5 for men and 3.6 for women; it was particularly elevated in younger patients [33]. In our study, the incidence of malignancies was at least 2.1 times higher in HIV-infected patients than in control patients (lower limit of the 95% confidence interval); however, the observed difference in NADM incidence between HIV-infected patients and control patients may not only be attributed to HIV-infection but also to individual life style factors such as smoking or alcohol consumption. This study has several limitations. Since this was a cohort study, patients were not evenly distributed with respect to cofactors associated with the outcomes of interest. However, the presence of certain cofactors may be attributed to the chronic disease itself or -in case of life style factors -may be attributed to the specific patient group. Therefore an even distribution of covariables may not be desirable as it would not have reflected the specific patient populations seen in clinical routine and would hence not have been appropriate for evaluating the disease burden in these populations. Gender and age distribution were brought into line during recruitment of patient groups (using frequency matching). For estimating associations between outcomes of interest and HIV-infection or diabetes mellitus, confounding was minimized by either stratification of analyses or by multivariable analyses.
One strength of this study lies in the majority of HIV-positive and HIV-negative patients seeking care at the same centers and therefore receiving similar screening methods, therapeutic options and treatment of comorbidities. Although the sample size was large, the study was not powered to detect differences in incident comorbidities. Apart from incident malignancies (the presence of which was excluded at baseline), we focused on prevalent diseases and disorders and were able to detect specific differences between groups. Of note, we did not adjust the level of significance for multiple testing and some outcomes only showed a weak association with HIV-infection. Concerning the DM group, factors associated with clinical outcomes -such as the duration of the disease and the indication for treatment -were not part of the assessment. Therefore, the DM group might have been a very heterogeneous group.
We were able to describe the burden of disease in three aging patient groups under medical treatment or medical observation. Aging antiretrovirally treated HIV-infected patients had a higher burden of comorbidity and symptoms than controls without severe chronic disease, but not than HIV-negative patients with diabetes mellitus. We hence emphasize the need for prevention strategies and good screening tools with regard to cardiovascular, renal, neurologic and malignant diseases in this patient population.