Obesity and Weight Gain Among HIV+ Uninsured Minorities
Obesity and Weight Gain Among HIV+ Uninsured Minorities
Of 1890 HIV+ patients receiving care in the HIV clinic from 2007 to 2010, 1214 met study inclusion criteria (Fig. 1). The study cohort was 62.8% Hispanic, 23.4% white, and 13.8% black. Only 7% had private insurance, 33% had Medicare, 14% had Medicaid, and 46% were uninsured. At the start of the observation period (hereafter referred to as baseline), 37.5% of the cohort was overweight and 22.1% obese (Table 1). Of 946 patients who were not obese at baseline, 112 (11.8%) became obese during follow-up. Median observation period was 3.25 years; 45 patients (3.7%) in the cohort had an observation period of 6 months to 1 year, the remainder had ≥1 year. Significant weight gain, defined as a ≥3% annual increase in BMI, was observed for 24.0% (n = 291) of the cohort. Baseline BMI categories for those patients who experienced significant weight gain were 148 (50.9%) normal weight, 95 (32.6%) overweight, and 48 (16.5%) obese.
All patient characteristics differed significantly by the 4-level race–ethnicity/insurance status variable except for CD4 category and virologic failure (Table 1). Uninsured minorities were younger, more likely to be women, and less likely to receive PI-based ART. Regardless of insurance type, minority patients were more likely to be overweight or obese [minority 62.4% vs. whites 50.4%; odds ratio (OR), 1.63; 95% CI: 1.25 to 2.13; P <0.001], and more likely to have diabetes (minority 16.8% vs. white 10.2%; OR, 1.77; CI: 1.16 to 2.70; P = 0.007). Insured patients were more likely to be overweight or obese at baseline (insured 57% vs. uninsured 43%; OR, 1.30; 95% CI: 1.03 to 1.64; P = 0.025). For the distribution of population characteristics by insurance category, see Table S1 (see Supplemental Digital Content, http://links.lww.com/QAI/A468).
In the unadjusted analysis, significant weight gain was more likely for uninsured whites (OR, 2.17; 95% CI: 1.18 to 3.98) and uninsured minorities (OR, 3.33; CI: 2.04 to 5.44) but insured minorities did not differ when compared with insured whites (Table 2). Patients with a normal baseline BMI had increased odds of significant weight gain compared with those who were obese at baseline, but overweight and obese patients did not differ significantly. Patients with severe immunosuppression (initial CD4 cell count <200 cells/μL) had greater odds of significant weight gain than patients with initial CD4 cell count ≥350 cells per microliter. Patients with virologic failure were less likely to have significant weight gain.
After adjustment for demographic and clinical variables, the adjusted odds ratio of significant weight gain was 2.85 (CI: 1.66 to 4.90) for uninsured minorities compared with whites with insurance but race–ethnicity/insurance groups did not differ significantly (Table 2). Persons with a normal baseline BMI had nearly 60% higher adjusted odds of significant weight gain compared with those who were obese at baseline. Severe immunosuppression was also associated with increased adjusted odds of significant weight gain than those with initial CD4 ≥350 cells per microliter.
A fully adjusted mixed-effects model including all BMI measurements revealed that uninsured minorities had a significantly more rapid increase in BMI over the observation time than other race–ethnicity/insurance categories (Fig 2; see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A468). For all 1214 patients in the cohort, the proportion of obese patients was projected to rise from 22.1% at baseline to 31.3% after 4 years. Of the 455 patients who were overweight at baseline, 131 (28.8%) were projected to become obese after 4 years (Table 3). The number of patients projected to be obese also varied by race–ethnicity/insurance category; the adjusted mixed-effects model revealed that 26% of insured whites, 22% of uninsured whites, 30% of insured minorities, and 38% of uninsured minorities were predicted to be obese at 4 years from baseline. Patients with severe immunosuppression and those without virologic failure had a significantly greater annual increases in BMI than those with less severe immunosuppression or with virologic failure (P < 0.001 and P = 0.04, respectively, see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A468). Compared with obese patients, patients in other BMI categories had significantly greater annual increases in BMI.
(Enlarge Image)
Figure 2.
Model-based least square mean BMI over time by race/ethnicity—insurance status. The least square mean BMIs were computed for a given race/insurance category at a given year using the final mixed effects model (see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A468) adjusting for gender, age, length of observation, virologic failure, CD4+ category, baseline BMI, ART, number of HIV clinic visits and years living with HIV. The observed margins are used for computation of least square means. For continuous variables, their mean values are used; and for categorical variables, the coefficients that are proportional to those observed in the data are used.
Results
Of 1890 HIV+ patients receiving care in the HIV clinic from 2007 to 2010, 1214 met study inclusion criteria (Fig. 1). The study cohort was 62.8% Hispanic, 23.4% white, and 13.8% black. Only 7% had private insurance, 33% had Medicare, 14% had Medicaid, and 46% were uninsured. At the start of the observation period (hereafter referred to as baseline), 37.5% of the cohort was overweight and 22.1% obese (Table 1). Of 946 patients who were not obese at baseline, 112 (11.8%) became obese during follow-up. Median observation period was 3.25 years; 45 patients (3.7%) in the cohort had an observation period of 6 months to 1 year, the remainder had ≥1 year. Significant weight gain, defined as a ≥3% annual increase in BMI, was observed for 24.0% (n = 291) of the cohort. Baseline BMI categories for those patients who experienced significant weight gain were 148 (50.9%) normal weight, 95 (32.6%) overweight, and 48 (16.5%) obese.
All patient characteristics differed significantly by the 4-level race–ethnicity/insurance status variable except for CD4 category and virologic failure (Table 1). Uninsured minorities were younger, more likely to be women, and less likely to receive PI-based ART. Regardless of insurance type, minority patients were more likely to be overweight or obese [minority 62.4% vs. whites 50.4%; odds ratio (OR), 1.63; 95% CI: 1.25 to 2.13; P <0.001], and more likely to have diabetes (minority 16.8% vs. white 10.2%; OR, 1.77; CI: 1.16 to 2.70; P = 0.007). Insured patients were more likely to be overweight or obese at baseline (insured 57% vs. uninsured 43%; OR, 1.30; 95% CI: 1.03 to 1.64; P = 0.025). For the distribution of population characteristics by insurance category, see Table S1 (see Supplemental Digital Content, http://links.lww.com/QAI/A468).
In the unadjusted analysis, significant weight gain was more likely for uninsured whites (OR, 2.17; 95% CI: 1.18 to 3.98) and uninsured minorities (OR, 3.33; CI: 2.04 to 5.44) but insured minorities did not differ when compared with insured whites (Table 2). Patients with a normal baseline BMI had increased odds of significant weight gain compared with those who were obese at baseline, but overweight and obese patients did not differ significantly. Patients with severe immunosuppression (initial CD4 cell count <200 cells/μL) had greater odds of significant weight gain than patients with initial CD4 cell count ≥350 cells per microliter. Patients with virologic failure were less likely to have significant weight gain.
After adjustment for demographic and clinical variables, the adjusted odds ratio of significant weight gain was 2.85 (CI: 1.66 to 4.90) for uninsured minorities compared with whites with insurance but race–ethnicity/insurance groups did not differ significantly (Table 2). Persons with a normal baseline BMI had nearly 60% higher adjusted odds of significant weight gain compared with those who were obese at baseline. Severe immunosuppression was also associated with increased adjusted odds of significant weight gain than those with initial CD4 ≥350 cells per microliter.
A fully adjusted mixed-effects model including all BMI measurements revealed that uninsured minorities had a significantly more rapid increase in BMI over the observation time than other race–ethnicity/insurance categories (Fig 2; see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A468). For all 1214 patients in the cohort, the proportion of obese patients was projected to rise from 22.1% at baseline to 31.3% after 4 years. Of the 455 patients who were overweight at baseline, 131 (28.8%) were projected to become obese after 4 years (Table 3). The number of patients projected to be obese also varied by race–ethnicity/insurance category; the adjusted mixed-effects model revealed that 26% of insured whites, 22% of uninsured whites, 30% of insured minorities, and 38% of uninsured minorities were predicted to be obese at 4 years from baseline. Patients with severe immunosuppression and those without virologic failure had a significantly greater annual increases in BMI than those with less severe immunosuppression or with virologic failure (P < 0.001 and P = 0.04, respectively, see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A468). Compared with obese patients, patients in other BMI categories had significantly greater annual increases in BMI.
(Enlarge Image)
Figure 2.
Model-based least square mean BMI over time by race/ethnicity—insurance status. The least square mean BMIs were computed for a given race/insurance category at a given year using the final mixed effects model (see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A468) adjusting for gender, age, length of observation, virologic failure, CD4+ category, baseline BMI, ART, number of HIV clinic visits and years living with HIV. The observed margins are used for computation of least square means. For continuous variables, their mean values are used; and for categorical variables, the coefficients that are proportional to those observed in the data are used.
Source...