Low bone mineral density among HIV-infected patients in Brazil

ABSTRACT

Decrease in bone mineral density (BMD) has been a complication among people living with HIV/AIDS. To investigate the prevalence of osteopenia/osteoporosis among HIV-infected people living in São Paulo city, we studied 108 HIV-infected patients (79 men and 29 women). We extracted data from patients’ medical records and BMD was measured by dual-energy X-ray absorptiometry (DXA). Median age of participants was 42 years (interquartile range [IQR] 36-48 years), and the median time since HIV diagnosis was 4.01 years (IQR 2-11 years). Patients had acquired HIV primarily by the sexual route (men who have sex with men 44%, heterosexual 49%). Median age, duration of HIV infection, duration of ART and CD4 nadir were similar for men and women. Plasma viral load was undetectable for 53 patients (49%). Median CD4 T cell count was 399 cells/µL (IQR 247 - 568). Twenty five patients (23%) had LBMD, and there was no statistically significant difference between men and women (<-1). The associated risk factors for LBMD were older age (≥ 50 years old) and smoking with a RR of 3.87 and 2.80, respectively. Thus, despite the lack of statistically significant relationship between the use of ART and LBMD or between duration of ART and LBMD, these factors should be addressed in larger studies.

INTRODUCTION

The HIV-infected population has aged, especially when engaged long-time survival, even in developing countries, such as Brazil, which have adopted universal access to ART since middle 90's ¹ . One of these new challenges is the chronic metabolic disturbances, such as diabetes, hypertension, metabolic syndrome and deregulation of bone metabolism ² .

The pathogenesis of bone demineralization in HIV infection is unclear ³ . Possible explanations include a direct action of the infection itself ⁴ , an untoward effect of the antiretroviral therapy (ART) ⁵ , a deregulation of bone metabolism as a result of the T-CD4 cell depletion ⁶ , or a consequence of other risk factors for osteoporosis that are more prevalent among people with HIV such as poor nutrition or smoking ⁷ . Identifying modifiable risk factors may be the key for designing simple interventions and preventing significant clinical morbidity.

In a middle-income country such as Brazil, bone demineralization is an important cause of morbidity, even in the absence of HIV infection. For example, from 2006 to 2008, 1% of all patients aging 60 years old and older who were hospitalized in the Unified National Health System (SUS) had a primary diagnosis of hip fracture, and they accounted for 2% of health care expenditures ⁸ . The mean cost of hospitalization for a hip fracture is $12,000 in the Brazilian private health care system ⁹ . Thus, as the HIV-infected population ages as a result of successful ART ^{10 , 11} , osteopenia and osteoporosis, both of them effects of HIV infection and aging, have the potential to become substantial public health burdens ¹² , and their prevention becomes an important public health goal. We studied the prevalence of low bone mineral density (LBMD), and risk factors in a population of 108 HIV-infected patients under follow-up at a referral center in São Paulo, Brazil.

MATERIAL AND METHODS

Subjects

The study population was composed of HIV-1-infected patients attended at the outpatient clinic for secondary immunodeficiencies of Hospital das Clínicas, University of São Paulo Medical School (HC/FMUSP), São Paulo, SP, Brazil, a large teaching and research hospital. This open patient cohort was composed of 455 active HIV1-infected subjects under follow-up for up to 16 years. Description of the cohort and its characteristics have been published elsewhere ^{13 , 14} . All patients were invited to participate, 108 accepted. The study protocol was approved by the Hospital das Clínicas Review Board and all patients gave written informed consent.

Predictor variables

Our predictor variables included sex, age, time since HIV diagnosis, CD4 cells/µL, CD4 nadir, plasma viral load, current use of ART, body mass index [BMI], current smoking status and physical activity. These variables were extracted from clinical charts. For analyzes, we dichotomized the age into ≤50 years and >50 years, CD4 count into <350 cells/µL and ≥350 cells/µL, CD4 nadir into <170 cells/µL and ≥170 cells/µL, time since HIV diagnosis into <4 years and ≥4 years, plasma viral load into detectable or non-detectable and time on ART as ≤5 years or >5 years. However, classes of ART drugs were not considered in detail on this report, therefore, no further information for their use is provided in the report.

Bone mineral density measurement and categorization

Our primary outcome variable was LBMD, which we measured using dual X-ray absorptiometry of the lumbar spine, femoral neck and total hip (Hologic, QDR 4500, Discovery model, Hologic Inc. Bedford, MA, USA). All measurements of bone mineral density (BMD) were performed by the same experienced technician (LT). We classified post-menopausal women and men ≥ 50 years old as having osteoporosis or osteopenia based on World Health Organization criteria. Subjects with a BMD T-score between -1 and -2.5 were categorized as having osteopenia, and subjects with a BMD T-score less than or equal to -2.5 were categorized as having osteoporosis. For the purposes of this study, any participant with osteopenia or osteoporosis according to the WHO definition wasconsidered to have LBMD. This classification has been extensively used by other authors ^{5 , 15 - 17} .

Statistical analysis

We calculated the prevalence of LBMD with 95% confidence intervals (CI), and evaluated the association of LBMD with its potential risk factors by calculating the relative risks.

RESULTS

We enrolled 108 patients of whom 78 (72%) were men and 30 (28%) were women. The median age was 43 years (interquartile range [IQR]: 43 - 48), and the median BMI was 24 kg/m² (IQR: 22 – 27). LBMD was detected in 15% (95% CI: 6 – 22) of patients younger than 50 years old and 54% (95% CI: 32 – 77) of those who were 50 years old and older. The median follow-up period from the date of HIV infection diagnosis was 4 years (95% CI: 2- 11); 74% of the patients were on ART for a median time of 5.12 years (IQR: 2.59 – 10.12). The median CD4 counts and median nadir for CD4 counts were 399 cells/µL (IQR: 275 – 566.5) and 176.5 cells/µL (IQR: 73 – 309), respectively. Participants enrolled in the study did not differ significantly from the remainder of patients followed in the Hospital das Clinicas cohort (Table 1).

Table 1. Demographic, clinical and behavioral characteristics according to low BMD, among 108 HIV-infected patients attending an outpatient HIV clinic, São Paulo, Brazil.

Variables		N= 108 patients
LBMD		25	23.15%
Gender
	Female	30	(27.78%)
	Male	78	72.22%
Age median (IQR)		43,17	43.17-48.19
	>50 Years	22	20.37%
	≤50 years	83	76.86%
HIV transmission
	Homosexual	50	46%
	Heterosexual	54	50%
	Others	4	3.70%
T CD4 cells Median (IQR)		399	275-566.5
	≤ 350 cells/mm3	45	41.67%
	≤ 350 cells/mm3	45	41.67%
Nadir median (IQR)		176,5	73-309
	<170 cells/mm3	50	46.30%
	≥170 cells/mm3	58	53.70%
Time since HIV diagnostic (IQR)		3,66	1.72-10.91
	>4 years	49	45.37%
	≤ 4 years	55	50.93%
Viral load
	Detectable	52	48.15%
	Undetectable	53	49.09%
Under ART
	Yes	80	74.07%
	No	28	25.93%
Years under ART
	>5 years	41	51.25%
	≤ 5 years	36	45%
Smoking
	Yes	28	25.93%
	No	78	72.22%
Physical Activity
	No	72	66.67%
	Yes	34	31.48%
BMI
	≤25	54	50%
	>25	35	32.41%

The overall prevalence of LBMD was 23.15% (95% CI, 15.06 – 31.23%), of them 15 (60%) HIV-1-infected had osteopenia and 10 (40%) HIV-1-infected had osteoporosis. The risk of LBMD was higher in subjects >50 years of age (14.45% vs. 54.54%, RR 3.77, 95% CI, 1.97 – 7.02, p=0.001), those with T CD4 counts <350 cells/µL (15.87% vs. 33.33%, RR 2.1, 95% CI 1.03-4.24, p=0.03) and current smokers (39.28% vs. 17.94%, RR 2.18, 95% CI 1.12 – 4.23, p=0.02) (Table 2).

Table 2. Risk factors for low BMD among HIV-infected patients attending an outpatient HIV clinic, São Paulo, Brazil.

Characteristic		All	LBMD present	LBMD not present	Probability of having LBMS given characteristic	RR	P Value
Gender
	Female	30 (27.8)	7 (28.0)	23 (27.6)	0.230	1 (0.46-2.12)	0.97
	Male	78 (72.2)	18 (72.0)	60 (72.3)	0.230
Age
	>50 years old	25 (23.2)	12 (50.00)	10 (12.4)	0.545	3.78 (1.97-7.02)	0.001
	≤50 years old	83 (76.9)	12 (50.0)	71 (87.7)	0.144
CD4+ T cells/µ
	≤ 350 CD4+ T cells/µ	45 (41.7)	15(60.0)	30 (36.1)	0.333	2.10 (1.03-4.24)	0.003
	>350 CD4+ T cells/µ	63 (58.3)	10 (40.0)	53 (63.9)	0.158
Nadir CD4
	<170 CD4+ T cells/µ	50 (46.3)	9 (36.0)	41 (49.4)	0.18	0.47 (0.31-1.43)	0.23
	≥170CD4+ T cells/µ	58 (53.7)	16 (64.0)	42 (50.6)	0.38
Time since HIV diagnosis
	>4 years	49 (47.1)	13 (54.2)	42 (52.5)	0.54	1.03 (0.46-1.92)	0.88
	≤ 4 years	55 (52.9)	11 (45.8)	38 (47.5)	0.52
Plasma HIV RNA
	Detectable	52 (49.5)	12 (50.0)	40 (49.4)	0.50	1.02 (0.50-2.05)	0.95
	Undetectabe	53 (50.5)	12 (50.0)	41 (50.6)	0.49
On ART
	Yes	80 (74.1)	21 (84.0)	59 (71.1)	0.84	1.18 (0.69-4.89)	0.19
	No	28 (25.9)	4 (16.0)	24 (28.9)	0.71
Years on ART
	>5	41 (52.6)	11 (55.0)	30 (51.7)	0.55	1.10 (0.51- 2.35)	0.20
	≤ 5	37 (47.4)	9 (45.0)	28 (48.2)	0.51
Smoking
	Yes	28 (26.4)	14 (56.0)	17 (21.0)	0.56	2.8 (1.12-4.23)	0.02
	No	78 (73.6)	11 (44.0)	64 (79.0)	0.20
Physical activity
	Not physically active	72 (67.9)	11 (44.0)	23 (71.6)	0.44	1.66 (0.84-3.27)	0.14
	Physically active	34 (32.1)	14 (56.0)	58 (28.4)	0.71
BMI, kg/m2 *
	≤25	54 (60.7)	18 (75.0)	36 (55.4)	0.75	1.94 (0.85-4.41)	0.09
	>25	35 (39.3)	6 (25.0)	29 (44.6)	0.55

LBMD, low bone mineral density; HIV, human immunodeficiency virus; RR, relative risk; CI, confidence interval; BMI, body mass index; ART, antiretroviral therapy;

^*19 missing values

DISCUSSION

We found that the overall prevalence of LBMD in our patients who had been on ART for a median of 5 years was 23.15%. Previous studies have estimated that the prevalence of osteopenia and osteoporosis in Brazilian patients with HIV infection was 54.5% ⁵ . This discrepancy could be related to different methodologies used in these studies, thus leading to these discordant results.

The CD4 cell count <350 cells/µL which was a risk factor for LBMD had not been previously reported and may indicate that the longer the duration of infection or the less successful the ART, the higher the likelihood of developing LBMD. Additionally, we did not find a statistically significant relationship between the use of ART and LBMD or between the duration of ART and LBMD, although both relationships have been already reported ^{18 , 19} . As mentioned before, osteoporosis was associated with HIV infection duration, low current or low nadir CD4 cell count or high viral load, for example ²⁰ . Cazanave et al. ²¹ showed that low nadir CD4 lymphocyte count was independently associated with osteoporosis/osteopenia in women, demonstrating that these alterations were induced by early bone demineralization.

In fact, HIV-infected infected subjects have lower BMI than HIV-negative control groups, as described previously ²² . Several reports demonstrated that habit of smoking, and lifestyle, traditional known risk factors for osteoporosis, are also important in this population ^{23 , 24} . Furthermore, reduced BMD was also significantly associated with low 25-hidroxy-calciferol in this cohort ²⁵ . Thus, smoking reduction, physical activity ²⁶ and other life changes may have an impact on bone disease.

Our study was limited by its small sample size, which may have led to a type II error. Larger and more regularly sampled cohorts may yield different results. Nonetheless, our findings are compelling and suggest that addressing modifiable risk factors for LBMD and fractures might constitute important interventions as Brazilian HIV-infected patients age, especially for those who are already older than 50 years old.