Supplement
Metabolic and Renal Adverse Effects of Antiretroviral Therapy in HIV-infected Children and Adolescents Clàudia Fortuny, MD, PhD,* Ángela Deyà-Martínez, MD,* Elena Chiappini, MD, PhD,† Luisa Galli, MD,† Maurizio de Martino, MD† and Antoni Noguera-Julian, MD, PhD*
Abstract: Worldwide, the benefits of combined antiretroviral (ARV) therapy in morbidity and mortality due to perinatally acquired human immunodeficiency virus infection are beyond question and outweigh the toxicity these drugs have been associated with in HIV-infected children and adolescents to date. In puberty, abnormal body fat distribution is stigmatizating and leads to low adherence to ARV treatment. The other metabolic comorbidities (mitochondrial toxicity, dyslipidemias, insulin resistance and low bone mineral density) and renal toxicity, albeit nonsymptomatic in most children, are increasingly being reported and potentially put this population at risk for early cardiovascular or cerebrovascular atherosclerotic disease, diabetes, pathologic fractures or premature renal failure in the third and fourth decades of life. Evidence from available studies is limited because of methodological limitations and also because of several HIV-unrelated factors influencing, to some degree, the development of these conditions. Current recommendations for the prevention, diagnosis, monitoring and treatment of metabolic and renal adverse effects in HIV-children and adolescents are based on adult studies, observational pediatric studies and experts’ consensus. Healthy lifestyle habits (regarding diet, exercise and refraining from toxic substances) and wise use of ARV options are the only preventive tools for the majority of patients. Should abnormal findings arise, switches in one or more ARV drugs have proved useful. Specific therapies are also available for some of these comorbidities, although the experience in the pediatric age is still very scarce. We aim to summarize the epidemiological, clinical and therapeutic aspects of metabolic and renal adverse effects in vertically HIV-infected children and adolescents. Key Words: antiretrovirals, human immunodeficiency virus, metabolic, nephropathy, pediatrics (Pediatr Infect Dis J 2015;34:S36–S43)
I
n the late 1990s, the implementation of highly active antiretroviral (ARV) therapies (HAART) dramatically changed the prognosis in vertically acquired HIV infection. The benefit obtained from HAART in terms of morbidity and mortality in HIV-infected children is beyond question, but ARV-related toxicities in the pediatric age have been increasingly reported over the past 15 years as well.1–9 Worldwide, current recommendations agree that all infants should be put on HAART as soon as HIV infection is diagnosed, leading to Accepted for publication November 16, 2014. From the *Infectious Disease Unit, Pediatrics Department, Hospital Sant Joan de Déu – Universitat de Barcelona, Barcelona, Spain; and †Paediatric Infectious Diseases Unit, Department of Paediatric Medicine, Anna Meyer Children’s University Hospital, University of Florence, Florence, Italy. This study was supported by Fundación para la Investigación y la Prevención del SIDA en España (FIPSE), Grants 36612/06 and 240813/09; and Fondo de Investigación Sanitaria (FIS, ISCIII), Grant Number 01738/13. Funding for this supplement was provided by the Italian Ministry of Health. The authors have no conflicts of interest to disclose. Address for correspondence: Antoni Noguera-Julian, MD, PhD, Infectious Disease Unit, Pediatrics Department, Hospital Sant Joan de Déu – Universitat de Barcelona, Passeig Sant Joan de Déu 2, 08950 Esplugues, Spain. E-mail: [email protected]. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/15/3405-0S36 DOI: 10.1097/INF.0000000000000663
S36 | www.pidj.com
potential lifelong exposure to these drugs and to the adverse effects associated with their use.10–12 Together with long-term metabolic complications, complex HAART regimens, insufficient pharmacokinetic and pharmacodynamic data and lack of appropriate formulations represent a great challenge for adherence to medications in HIV-infected children, and ultimately lead to viral failure and emergence of drug-resistant viral mutations. ARV-associated metabolic complications were initially documented in adult patients, but all of them have also been reported in the pediatric age afterwards. These include mitochondrial toxicity, lipodystrophy (LD) syndrome (abnormal body fat distribution), dyslipidemias, insulin resistance, cardiovascular and cerebrovascular risk, and low bone mineral density (BMD); renal toxicity is also becoming more and more prevalent in the HIV-infected pediatric population, due to the increasing use of tenofovir.13 In the late 1990s, HIV-infected teenagers from America and Europe were predominantly affected in the first case series and cohort studies. Later on, and because of the widespread availability of ARV therapy, metabolic adverse events have also been described in resource-limited settings3,4,7–9,14 and in younger children.4,7,8,15 Although exposure to certain ARVs plays a central role in the development of these clinical syndromes, most of them have also been reported in the naïve patient. This highlights the deleterious effect of chronic inflammatory states due to uncontrolled HIV infection, also in the development of metabolic comorbidities. In the pediatric age, available studies are limited by the number of patients, the observational nature of the studies, and the lack of long-term follow-up and appropriate control groups of HIV-uninfected peers. Ethnic background and different stages of growth and development, such as puberty, are also known to have an influence on the development/worsening of some of these adverse events, as well as some other HIV-unrelated factors (ie, malnutrition in resource-limited settings, the obesity epidemic, coinfections, the degree of solar exposure and genetic background). Finally, differences in definitions and methods used to determine some of these conditions render some results unreliable and comparison between studies very complex. We aim to summarize the epidemiological, clinical and therapeutic aspects of metabolic and renal adverse effects in vertically HIV-infected children and adolescents, as well as the potential long-term morbidity these confer on this vulnerable population growing into adulthood.
MITOCHONDRIAL TOXICITY Mitochondrial alterations have been extensively associated with HIV infection16 and with ARV drugs in adulthood and children.17–19 Nucleoside reverse transcriptase inhibitors are known to inhibit DNA polymerase gamma,20 but protease inhibitors have also been described as causing mitochondrial apoptosis.21 Many of the metabolic adverse events described in HIV-infected persons are believed to be due, at least partially, to mitochondrial toxicity resulting from inhibition of mitochondrial DNA polymerase gamma and secondary depletion of mitochondrial DNA in fat, muscle, peripheral blood mononuclear cells and other tissues.22–24
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
In HIV-infected children, chronic symptom-free hyperlactatemia (normal values up to 2.1 mmol/L) has been reported in up to one-third of patients in some cohorts,25,26 but symptomatic hyperlactatemia with or without lactic acidosis has seldom been observed.27–31 Of note, mitochondrial dysfunction, hyperlactatemia and several symptomatic cases have also been described in HIVuninfected ARV-exposed healthy infants.32–34 Sporadic cases of lactic acidosis have been reported with all available nucleoside analogs, but exposure to stavudine and didanosine is associated with the highest risk, especially when the two drugs are used together; in fact, the use of stavudine is discouraged in the latest treatment guidelines, whereas didanosine is no longer considered a first-line option.11 Other risk factors that have been described in the adult patient are advanced HIV infection, AfricanAmerican race, hepatitis C virus coinfection, pregnancy, obesity and coadministration of other drugs with either stavudine or didanosine (including metformin, tenofovir and ribavirin).11,35 The clinical presentation of lactic acidosis is unspecific and may include gastrointestinal symptoms (nausea and vomiting, abdominal pain), dyspnea, peripheral neuropathy and systemic symptoms (fatigue, weight loss and myalgias), developing over several days or presenting as fulminant multi-organ failure; hepatic and pancreatic involvement is often observed as well. The diagnosis of lactic acidosis is by exclusion and other conditions potentially leading to secondary acidosis (ie, sepsis, diarrhea, etc.) have to be ruled out previously; thereafter, elevated blood lactate levels (usually >5 mmol/L) together with elevated anion gap metabolic acidosis will confirm the diagnosis. In the absence of symptoms, the routine measurement of serum lactate is not recommended, because of its low specificity. Lactic acidosis is probably the only metabolic adverse event in which immediate discontinuation of all ARV agents is mandatory. Early diagnosis and implementation of supportive therapy (intravenous fluids, oxygen, etc.) is critical, because lactic acidosis is associated with high mortality rates in the adult patient. Together with supportive therapy, supplementation with vitamins, carnitine or co-enzyme Q10 has been recommended by some experts. A nucleoside-sparing regimen or a HAART combination including mitochondrion-friendly nucleosides (ie, abacavir, tenofovir and lamivudine/emtricitabine) is recommended for resuming treatment in patients previously affected with an episode of symptomatic hyperlactatemia; close clinical follow-up and monthly monitoring of lactate in the first months of the new HAART regimen is highly recommended in these cases.
LD SYNDROME (ABNORMAL BODY FAT DISTRIBUTION) Classically, 3 different clinical patterns have been described for LD: (1) lipoatrophy, characterized by thinning of subcutaneous
ARV-related Comorbidities in HIV+ Children
fat in face, buttocks and extremities, which makes veins look more prominent; (2) lipohypertrophy or central fat accumulation, which associates increased abdominal girth, dorsocervical fat pad (socalled buffalo hump) and breast hypertrophy in girls, and gynecomastia in boys; and (3) a mixed pattern combining both lipoatrophy and lipohypertrophy. Abnormal body fat distribution is commonly associated with dyslipidemias and insulin resistance, and it often becomes stigmatizating, especially among female adolescents, and leads to poor adherence and treatment failure. Although stavudine (and also zidovudine) use has been clearly blamed for the development of lipoatrophy, several risk factors have been identified in the case of lipohypertrophy, including pubertal development, older age and longer time on HAART, HIV disease, previous obesity, a sedentary lifestyle, a genetic predisposition and exposure to protease inhibitors and efavirenz. Highly variable prevalence of LD has been reported in different cohorts, ranging from 1% to 65% (Table 1), each clinical pattern affecting approximately one-third of patients. In those settings, where stavudine-containing HAART regimens are still common, lipoatrophy is predominant.7,9,36 Clinical identification of the characteristic body-shape changes by the physician, caretaker or even the patient himself remains the most commonly used method for the diagnosis of LD. Anthropometric measurements (waist-to-hip ratio, skinfolds and limb circumference) only measure subcutaneous fat, and they are reproducible and cheap, but they lack standardization. Image techniques (computed tomography and magnetic resonance imaging) discriminate between visceral and subcutaneous fat. Finally, dualenergy X-ray absorptiometry (DXA) also allows the differentiation of visceral and subcutaneous fat, assesses BMD, exhibits low radiation doses, and has been used in longitudinal studies; good correlations have been observed between patient and physicianrated lipoatrophy scores and DXA scans.37 Unfortunately, both image techniques and DXA are expensive and often unavailable in resource-limited settings. In the absence of sufficient evidence for pharmacological treatments and other medical strategies, recommendations regarding lifestyle modifications promoting healthy diet and exercise, and avoiding smoking and other toxic substances, are probably the most important therapeutic approach to prevention and treatment of metabolic complications, including LD, in the HIV-infected pediatric population (Table 2a).10–12,35 In the case of lipoatrophy, avoiding the use of thymidine analogs (especially stavudine, but also zidovudine) is recommended. Some switch studies have been published, most commonly replacing stavudine with another nucleoside analog, and they have reported either a halt in lipoatrophy progression38 or significant clinical improvements.39,40 Conversely, switch strategies for the management of central fat accumulation, generally replacing a
TABLE 1. Summary of Different Cohort Studies Describing the Prevalence Rate of Abnormal Fat Distribution, Dyslipidemias and Insulin Resistance in HIV-infected Children and Adolescents
References EPLG1 Rosso et al2 Aurpibul et al3 Brewinski et al4 Dapena et al5 Alam et al6 Piloya et al7 Hazra et al8 Kinabo et al9
Number of Patients, Region of Origin
Median Age, Median Time on HAART
Prevalence of LD (LA, LH and MP)
Prevalence of Elevated TG
Prevalence of Elevated TC
n = 477, Europe n = 48, Italy n = 90, Thailand n = 477, Latin America n = 157, Spain n = 426, Europe n = 364, Uganda n = 249, South America n = 210, Tanzania
9.8 yrs, 4.7 yrs 12 yrs, 5.1 yrs 10.6 yrs, 3yrs 6.8 yrs, 2.4 yrs 13 yrs, 10.2 yrs 12.2 yrs, 5.2 yrs 8 yrs, 3.8 yrs 7.1 yrs (mean), 5.6 yrs 9.6 yrs, NR
26% (7.5, 8.8 and 6.4%) 33% (8, 6 and 19%) 65% (13, 29.9 and 22.1%) NR 40.5% (20.4, 5.1 and 14.6%) 42% (28 and 27%; MP NR) 27% (22 and 4.7%; MP NR) NR 30% (19, 3.8 and 7.1%)
21% 14.6% 12% 29.4% 24.8% 16.7% 28.6% 33.6% NR
27% 16.7% 11% 20.5% 23.9% 12.4% 5.8% 12.7% NR
Prevalence of Insulin Resistance NR 52% 4% NR 20% 1.6% 0% 6.8% NR
LA indicates lipoatrophy; LH, lipohypertrophy; MP, mixed pattern; TC, total cholesterol; NR, not reported.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pidj.com | S37
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
Fortuny et al
TABLE 2. (a) Recommendations Regarding Healthy Lifestyle Habits; and (b) Monitoring Proposal for Metabolic and Renal Adverse Effects in the Routine Follow-up of HIV-infected Children and Adolescents (a) Calorically appropriate low-fat diet, especially in patients with high BMI or dyslipidemias Ensure sufficient calcium and vitamin D intake Regular aerobic exercise; weight-bearing exercise from puberty Smoking avoidance or cessation Avoidance of other toxic substances, illicit or not (b) Body weight and height at each visit, calculate BMI; blood pressure measurement at least every 3–6 months Clinical assessment of metabolic toxicities at each visit, including abnormal body fat distribution patterns In the absence of symptoms, lactate determination is not recommended Laboratory monitoring for dyslipidemias and glucose homeostasis: before and early after (4–12 weeks) HAART initiation or switch, and at least every 6–12 months thereafter if no ongoing toxicity concerns Periodic measurement of vitamin D, especially when several risk factors for low BMD coincide If available, DXA scans should be performed periodically, every 1–2 years, especially among children with abnormal findings and/or when several risk factors for low BMD coincide Urinalysis (proteinuria and glucosuria) and calculate estimated glomerular filtration rate at least every 6–12 months; every 3–6 months if tenofovir is used
protease inhibitor with either nevirapine or efavirenz, have shown disappointing results to date.38,41,42 Although the data are still insufficient, corrective surgery of severe facial lipoatrophy has been successfully used in teenagers43 and has become widely used in the adult patient, for central fat accumulation as well as facial.44 Other therapeutic possibilities should not be considered in children out of clinical trials (ie, recombinant growth hormone, testosterone, leptin, metformin and thiazolinediones).
DYSLIPIDEMIAS Normal lipid values for children are defined according to percentile cutpoints within population distributions. It should be kept in mind that age, sex and race have significant effects on lipid levels, raising doubts about the utility of fixed screening cutpoints. Twelve-hour fasting lipid level cutpoints commonly used for HIV infection in the pediatric age are as follows: triglyceride (TG) >100 mg/dL up to 10 years of age and >130 mg/dL in older patients, total cholesterol >200 mg/dL, low-density lipoprotein >130 mg/dL and high-density lipoprotein ≤40 mg/dL.45 Lipid abnormalities were described early in the HIVinfected pediatric population and have been associated mainly with exposure to protease inhibitors (especially with ritonavir boosting), stavudine and efavirenz, fat maldistribution and HIV disease,46–48 even in very young infants.15 HIV-unrelated factors linked to dyslipidemias have also been identified in this population, such as family history, inadequate diet, sedentarism, obesity, hypertension and smoking. Dyslipidemias may develop a few weeks after beginning therapy and are not associated with symptoms, although they will probably play a central role in future atherosclerotic disease. Again, wide ranges in prevalence of dyslipidemias have been reported in HIV-infected pediatric patients. In 2005, the PACTG 219C Study Group reported a 13% prevalence of hyperTC in the largest cohort of HIV-infected children and adolescents analyzed to date;46 later on, the same group described a resolution to normal total cholesterol within 2 years of follow-up in one-third of their patients.49 Other study figures for hyperTC and hyperTG range from 5.8% to 27% and from 12% to 33.6%, respectively (Table 1). Adolescents living with HIV infection should undergo regular overnight fasting lipid panel monitoring when initiating or changing HAART and at least every 6–12 months thereafter, together with routine blood analysis. In younger children, regular monitoring can be based on nonfasting samples; however, it should be kept in mind that nonfasting falsely elevated TG and low-density lipoprotein levels can be observed.
S38 | www.pidj.com
The same recommendations on lifestyle habits apply to the prevention of dyslipidemia (Table 2a), together with assessment of additional cardiovascular risk factors. Besides these, children with elevated TG levels should follow a diet with low simple carbohydrates. Some studies in children have reported improvements in lipid results upon switching from a protease inhibitor to a nonnucleoside.41,50–52 When possible, lipid-friendly ARVs should be chosen for treatment; in this respect, darunavir and atazanavir, even when ritonavir-boosted, as well as new ARV drug classes, have shown better lipid effects than older protease inhibitors in adult studies. When dyslipidemia occurs, a proper 3- to 6-month trial period of lifestyle modifications is recommended before any change in HAART or pharmacologic treatment is implemented. Any new treatment risk must be balanced against potential benefits (pharmacokinetic interactions, adverse events, pill burden, etc.). Lipid-lowering therapies should always be managed together with a lipid specialist and according to specific guidelines.10,11,35 Statins (pravastatin, atorvastatin, fluvastatin and rosuvastatin) are used to lower cholesterol levels by inhibiting the enzyme HMG-CoA reductase and some of them are approved for children aged 10 years or more. Side effects of statins include muscle pain, and increased risk of diabetes and hepatitis. All of them but pravastatin show multiple drug interactions with ARV. Ezetimibe acts by decreasing cholesterol absorption in the small intestine, and it is used alone or as add-on therapy with statins; it is also available for adolescents. Fibrates and omega-3 fatty acids are used in adults as secondary therapies of hyperTG but are not approved for the pediatric age. To date, the use of these drugs in the setting of pediatric HIV infection has been very scarcely reported.49,53
INSULIN RESISTANCE In contrast to adults, in children and youth living with HIV the spectrum of insulin resistance, asymptomatic hyperglycemia and diabetes mellitus appears to be a relatively uncommon finding. Insulin resistance is defined by means of elevated insulin levels or with the homeostatic model of insulin resistance (HOMA-IR = fasting insulin (μIU/mL) × fasting glucose (mg/dL)/405); cutoffs of 2.5 and 4.0 for HOMA-IR values are used in Tanner 1 and Tanner 2 and higher subjects, respectively. The prevalence of insulin resistance in different cohorts ranges from 0% to 52%, but elevated fasting hyperglycemia (>100 mg/dL) has only rarely been reported (Table 1).8,54–58 Several risk factors for the development of insulin resistance have been identified, with ARV use (mainly protease inhibitors and thymidine analogues) and elevated body mass index © 2015 Wolters Kluwer Health, Inc. All rights reserved.
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
being the most important, but also fat maldistribution, advanced HIV disease, hepatitis C virus coinfection, family history, dyslipidemias, older age and puberty.2,54–56,59 To the best of our knowledge, no cases of frank diabetes mellitus type 2 have been reported in HIV-infected patients in the pediatric age, but typical symptoms to bear in mind are polydipsia, polyuria, polyphagia and changes in body habitus. Acanthosis nigricans, a nonsymptomatic skin condition due to hyperinsulinemia that causes one or more areas of skin to darken, thicken and feel velvety, seldom occurs in this population.60 In our experience, complete resolution of A. nigricans is observed upon protease inhibitor cessation. In the follow-up of impaired glucose metabolism in HIVinfected pediatric patients, only random or overnight fasting plasma glucose levels are recommended at HAART initiation or switch and every 6–12 months thereafter. Whenever random/fasting glucose levels over 140/100 mg/dL are observed or symptoms arise, referral to a pediatric endocrinologist is advisable. Determination of insulin levels or glycated hemoglobin levels, and glucose tolerance tests, are not routinely recommended. Primary prevention and initial treatment of glucose metabolism disturbances include lifestyle modifications (Table 2a) and avoidance or switching of certain ARV drugs (stavudine, indinavir, etc). An observational switch study in HIV-infected children showed improvements in glucose homeostasis parameters following protease inhibitor replacement.61 The use of metformin (licensed in children aged 10 years or more) and thiazolindinediones (rosiglitazone and pioglitazone; not approved below the age of 18 years) has not been reported in HIV-infected children to date. The former should be used with caution, because of the risk of lactic acidosis.
CARDIOVASCULAR AND CEREBROVASCULAR RISK HIV-infected adult patients are at a higher risk of early cardiovascular and cerebrovascular disease than the general population, due to HIV infection itself and also to ARV use.62,63 In recent years, there has been growing evidence of an increased global risk of premature cardiovascular disease also in children and adolescents with vertically acquired HIV infection. The overweight/obesity epidemic due to lifestyle factors among youth is also affecting the HIV-infected pediatric population, with prevalence rates up to 8.4% in Spain5 and 12% in the United States.64 Together with obesity and other HIV-unrelated factors, abnormal body fat distribution, lipid abnormalities and altered glucose metabolism lead to increased risk of future cardiovascular disease in children.65 For instance, excess trunk fat in healthy children was associated with elevated blood pressure levels, lipid levels and insulin levels in the Bogalusa Heart Study.66,67 Finally, uncontrolled HIV infectionrelated inflammatory and immunologic factors also contribute to cardiovascular risk.68 Recently, by means of the Pathobiological Determinants of Atherosclerosis in Youth (PDAY) scoring system,69 Patel et al70 described an increased coronary/abdominal aorta disease risk in 48%/24% of a large cohort of HIV-infected adolescents in the United States, with the highest scores being predicted by HIV disease severity and boosted protease inhibitor use. The cardiovascular toxicity associated with abacavir use remains controversial. In line with the former, several authors have reported increased carotid intima-media thickness and arterial stiffness as assessed by pulse wave velocity among HIV-infected children and adolescents when compared with healthy controls, often together with elevated metabolic and inflammatory markers of atherosclerotic disease, such as high-sensitive C-reactive protein, lipid levels, glycated hemoglobin and activated T cells.71–74 Miller et al75 also reported differences in markers of coagulant and endothelial dysfunction. Finally, more © 2015 Wolters Kluwer Health, Inc. All rights reserved.
ARV-related Comorbidities in HIV+ Children
than 50% of a small cohort of mostly vertically HIV-infected young adults showed coronary artery abnormalities consisting of luminal narrowing and irregularity of the coronary vessel wall with cardiac magnetic resonance techniques.76 Fortunately, cerebrovascular and cardiovascular clinical events in perinatally HIV-infected patients are still rare, although an increase in incidence is foreseeable in the coming decades.
LOW BMD AND OSTEOPOROSIS Low BMD is a recognized metabolic complication of HIV disease that has been noted more frequently in the era of HAART.77 A meta-analysis demonstrated a 15% prevalence of osteoporosis and a 67% prevalence of low BMD in HIV-infected adults some years ago;77 recently, a modest increase in incident fractures was also associated with HIV infection.78 Although low BMD has also been described in naïve patients, exposure to protease inhibitor-based HAART regimens and especially to tenofovir was associated with higher rates of osteoporosis when compared with other drugs.79,80 During childhood and adolescence, bone mass increases extensively through longitudinal growth. Several factors influence the absolute amount of peak bone mass attained, including gender, ethnicity, genetics, weight-bearing physical activity, dietary intake of calcium and vitamin D and pubertal development through hormonal influence. Peak bone mass occurs around late adolescence and, when suboptimal, becomes an important predictor of the future risk of osteoporosis and fractures in adulthood. DXA is the preferred method for assessing BMD in children and adolescents, and the spine and total body are the preferred skeletal sites. DXA values in children should always be adjusted using Z score height, especially in patients with growth delay, such as some HIV-infected patients, and they must be compared with an appropriate reference data set. In the pediatric age, the diagnosis of osteoporosis should not be made on the basis of densitometric criteria alone; the presence of vertebral compression (crush) fractures or low-energy peripheral long bones fractures is mandatory, together with a BMD Z score equal or below -2.0 (http://www.iscd. org/2013-iscd-official-positions-pediatric/). A BMD Z score equal to or below -2.0 alone defines low BMD in children; the term osteopenia should not be used. In vertically infected HIV-positive children and adolescents, several conditions associated with decreased BMD may coexist together with HIV infection, such as lack of mobility, prematurity, delayed puberty and growth and low body mass. Low BMD values have been consistently reported among different HIV-infected pediatric cohorts when compared with HIV-uninfected peers throughout the world, with prevalence rates ranging from 7% to 32%.81–84 In these studies, time on HAART, use of tenofovir, stavudine and protease inhibitors, advanced HIV disease and late pubertal stages were identified as risk factors for low BMD,81–85 whereas vitamin D status was not.81–83 Importantly, low BMD was also described among prepubertal patients,84 calling attention to the high risk of sub-optimal peak BMD should this situation remain unchanged through puberty. Vitamin D plays a central role in bone growth and remodeling. Highly variable rates of hypovitaminosis D (
Metabolic and Renal Adverse Effects of Antiretroviral Therapy in HIV-infected Children and Adolescents Clàudia Fortuny, MD, PhD,* Ángela Deyà-Martínez, MD,* Elena Chiappini, MD, PhD,† Luisa Galli, MD,† Maurizio de Martino, MD† and Antoni Noguera-Julian, MD, PhD*
Abstract: Worldwide, the benefits of combined antiretroviral (ARV) therapy in morbidity and mortality due to perinatally acquired human immunodeficiency virus infection are beyond question and outweigh the toxicity these drugs have been associated with in HIV-infected children and adolescents to date. In puberty, abnormal body fat distribution is stigmatizating and leads to low adherence to ARV treatment. The other metabolic comorbidities (mitochondrial toxicity, dyslipidemias, insulin resistance and low bone mineral density) and renal toxicity, albeit nonsymptomatic in most children, are increasingly being reported and potentially put this population at risk for early cardiovascular or cerebrovascular atherosclerotic disease, diabetes, pathologic fractures or premature renal failure in the third and fourth decades of life. Evidence from available studies is limited because of methodological limitations and also because of several HIV-unrelated factors influencing, to some degree, the development of these conditions. Current recommendations for the prevention, diagnosis, monitoring and treatment of metabolic and renal adverse effects in HIV-children and adolescents are based on adult studies, observational pediatric studies and experts’ consensus. Healthy lifestyle habits (regarding diet, exercise and refraining from toxic substances) and wise use of ARV options are the only preventive tools for the majority of patients. Should abnormal findings arise, switches in one or more ARV drugs have proved useful. Specific therapies are also available for some of these comorbidities, although the experience in the pediatric age is still very scarce. We aim to summarize the epidemiological, clinical and therapeutic aspects of metabolic and renal adverse effects in vertically HIV-infected children and adolescents. Key Words: antiretrovirals, human immunodeficiency virus, metabolic, nephropathy, pediatrics (Pediatr Infect Dis J 2015;34:S36–S43)
I
n the late 1990s, the implementation of highly active antiretroviral (ARV) therapies (HAART) dramatically changed the prognosis in vertically acquired HIV infection. The benefit obtained from HAART in terms of morbidity and mortality in HIV-infected children is beyond question, but ARV-related toxicities in the pediatric age have been increasingly reported over the past 15 years as well.1–9 Worldwide, current recommendations agree that all infants should be put on HAART as soon as HIV infection is diagnosed, leading to Accepted for publication November 16, 2014. From the *Infectious Disease Unit, Pediatrics Department, Hospital Sant Joan de Déu – Universitat de Barcelona, Barcelona, Spain; and †Paediatric Infectious Diseases Unit, Department of Paediatric Medicine, Anna Meyer Children’s University Hospital, University of Florence, Florence, Italy. This study was supported by Fundación para la Investigación y la Prevención del SIDA en España (FIPSE), Grants 36612/06 and 240813/09; and Fondo de Investigación Sanitaria (FIS, ISCIII), Grant Number 01738/13. Funding for this supplement was provided by the Italian Ministry of Health. The authors have no conflicts of interest to disclose. Address for correspondence: Antoni Noguera-Julian, MD, PhD, Infectious Disease Unit, Pediatrics Department, Hospital Sant Joan de Déu – Universitat de Barcelona, Passeig Sant Joan de Déu 2, 08950 Esplugues, Spain. E-mail: [email protected]. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/15/3405-0S36 DOI: 10.1097/INF.0000000000000663
S36 | www.pidj.com
potential lifelong exposure to these drugs and to the adverse effects associated with their use.10–12 Together with long-term metabolic complications, complex HAART regimens, insufficient pharmacokinetic and pharmacodynamic data and lack of appropriate formulations represent a great challenge for adherence to medications in HIV-infected children, and ultimately lead to viral failure and emergence of drug-resistant viral mutations. ARV-associated metabolic complications were initially documented in adult patients, but all of them have also been reported in the pediatric age afterwards. These include mitochondrial toxicity, lipodystrophy (LD) syndrome (abnormal body fat distribution), dyslipidemias, insulin resistance, cardiovascular and cerebrovascular risk, and low bone mineral density (BMD); renal toxicity is also becoming more and more prevalent in the HIV-infected pediatric population, due to the increasing use of tenofovir.13 In the late 1990s, HIV-infected teenagers from America and Europe were predominantly affected in the first case series and cohort studies. Later on, and because of the widespread availability of ARV therapy, metabolic adverse events have also been described in resource-limited settings3,4,7–9,14 and in younger children.4,7,8,15 Although exposure to certain ARVs plays a central role in the development of these clinical syndromes, most of them have also been reported in the naïve patient. This highlights the deleterious effect of chronic inflammatory states due to uncontrolled HIV infection, also in the development of metabolic comorbidities. In the pediatric age, available studies are limited by the number of patients, the observational nature of the studies, and the lack of long-term follow-up and appropriate control groups of HIV-uninfected peers. Ethnic background and different stages of growth and development, such as puberty, are also known to have an influence on the development/worsening of some of these adverse events, as well as some other HIV-unrelated factors (ie, malnutrition in resource-limited settings, the obesity epidemic, coinfections, the degree of solar exposure and genetic background). Finally, differences in definitions and methods used to determine some of these conditions render some results unreliable and comparison between studies very complex. We aim to summarize the epidemiological, clinical and therapeutic aspects of metabolic and renal adverse effects in vertically HIV-infected children and adolescents, as well as the potential long-term morbidity these confer on this vulnerable population growing into adulthood.
MITOCHONDRIAL TOXICITY Mitochondrial alterations have been extensively associated with HIV infection16 and with ARV drugs in adulthood and children.17–19 Nucleoside reverse transcriptase inhibitors are known to inhibit DNA polymerase gamma,20 but protease inhibitors have also been described as causing mitochondrial apoptosis.21 Many of the metabolic adverse events described in HIV-infected persons are believed to be due, at least partially, to mitochondrial toxicity resulting from inhibition of mitochondrial DNA polymerase gamma and secondary depletion of mitochondrial DNA in fat, muscle, peripheral blood mononuclear cells and other tissues.22–24
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
In HIV-infected children, chronic symptom-free hyperlactatemia (normal values up to 2.1 mmol/L) has been reported in up to one-third of patients in some cohorts,25,26 but symptomatic hyperlactatemia with or without lactic acidosis has seldom been observed.27–31 Of note, mitochondrial dysfunction, hyperlactatemia and several symptomatic cases have also been described in HIVuninfected ARV-exposed healthy infants.32–34 Sporadic cases of lactic acidosis have been reported with all available nucleoside analogs, but exposure to stavudine and didanosine is associated with the highest risk, especially when the two drugs are used together; in fact, the use of stavudine is discouraged in the latest treatment guidelines, whereas didanosine is no longer considered a first-line option.11 Other risk factors that have been described in the adult patient are advanced HIV infection, AfricanAmerican race, hepatitis C virus coinfection, pregnancy, obesity and coadministration of other drugs with either stavudine or didanosine (including metformin, tenofovir and ribavirin).11,35 The clinical presentation of lactic acidosis is unspecific and may include gastrointestinal symptoms (nausea and vomiting, abdominal pain), dyspnea, peripheral neuropathy and systemic symptoms (fatigue, weight loss and myalgias), developing over several days or presenting as fulminant multi-organ failure; hepatic and pancreatic involvement is often observed as well. The diagnosis of lactic acidosis is by exclusion and other conditions potentially leading to secondary acidosis (ie, sepsis, diarrhea, etc.) have to be ruled out previously; thereafter, elevated blood lactate levels (usually >5 mmol/L) together with elevated anion gap metabolic acidosis will confirm the diagnosis. In the absence of symptoms, the routine measurement of serum lactate is not recommended, because of its low specificity. Lactic acidosis is probably the only metabolic adverse event in which immediate discontinuation of all ARV agents is mandatory. Early diagnosis and implementation of supportive therapy (intravenous fluids, oxygen, etc.) is critical, because lactic acidosis is associated with high mortality rates in the adult patient. Together with supportive therapy, supplementation with vitamins, carnitine or co-enzyme Q10 has been recommended by some experts. A nucleoside-sparing regimen or a HAART combination including mitochondrion-friendly nucleosides (ie, abacavir, tenofovir and lamivudine/emtricitabine) is recommended for resuming treatment in patients previously affected with an episode of symptomatic hyperlactatemia; close clinical follow-up and monthly monitoring of lactate in the first months of the new HAART regimen is highly recommended in these cases.
LD SYNDROME (ABNORMAL BODY FAT DISTRIBUTION) Classically, 3 different clinical patterns have been described for LD: (1) lipoatrophy, characterized by thinning of subcutaneous
ARV-related Comorbidities in HIV+ Children
fat in face, buttocks and extremities, which makes veins look more prominent; (2) lipohypertrophy or central fat accumulation, which associates increased abdominal girth, dorsocervical fat pad (socalled buffalo hump) and breast hypertrophy in girls, and gynecomastia in boys; and (3) a mixed pattern combining both lipoatrophy and lipohypertrophy. Abnormal body fat distribution is commonly associated with dyslipidemias and insulin resistance, and it often becomes stigmatizating, especially among female adolescents, and leads to poor adherence and treatment failure. Although stavudine (and also zidovudine) use has been clearly blamed for the development of lipoatrophy, several risk factors have been identified in the case of lipohypertrophy, including pubertal development, older age and longer time on HAART, HIV disease, previous obesity, a sedentary lifestyle, a genetic predisposition and exposure to protease inhibitors and efavirenz. Highly variable prevalence of LD has been reported in different cohorts, ranging from 1% to 65% (Table 1), each clinical pattern affecting approximately one-third of patients. In those settings, where stavudine-containing HAART regimens are still common, lipoatrophy is predominant.7,9,36 Clinical identification of the characteristic body-shape changes by the physician, caretaker or even the patient himself remains the most commonly used method for the diagnosis of LD. Anthropometric measurements (waist-to-hip ratio, skinfolds and limb circumference) only measure subcutaneous fat, and they are reproducible and cheap, but they lack standardization. Image techniques (computed tomography and magnetic resonance imaging) discriminate between visceral and subcutaneous fat. Finally, dualenergy X-ray absorptiometry (DXA) also allows the differentiation of visceral and subcutaneous fat, assesses BMD, exhibits low radiation doses, and has been used in longitudinal studies; good correlations have been observed between patient and physicianrated lipoatrophy scores and DXA scans.37 Unfortunately, both image techniques and DXA are expensive and often unavailable in resource-limited settings. In the absence of sufficient evidence for pharmacological treatments and other medical strategies, recommendations regarding lifestyle modifications promoting healthy diet and exercise, and avoiding smoking and other toxic substances, are probably the most important therapeutic approach to prevention and treatment of metabolic complications, including LD, in the HIV-infected pediatric population (Table 2a).10–12,35 In the case of lipoatrophy, avoiding the use of thymidine analogs (especially stavudine, but also zidovudine) is recommended. Some switch studies have been published, most commonly replacing stavudine with another nucleoside analog, and they have reported either a halt in lipoatrophy progression38 or significant clinical improvements.39,40 Conversely, switch strategies for the management of central fat accumulation, generally replacing a
TABLE 1. Summary of Different Cohort Studies Describing the Prevalence Rate of Abnormal Fat Distribution, Dyslipidemias and Insulin Resistance in HIV-infected Children and Adolescents
References EPLG1 Rosso et al2 Aurpibul et al3 Brewinski et al4 Dapena et al5 Alam et al6 Piloya et al7 Hazra et al8 Kinabo et al9
Number of Patients, Region of Origin
Median Age, Median Time on HAART
Prevalence of LD (LA, LH and MP)
Prevalence of Elevated TG
Prevalence of Elevated TC
n = 477, Europe n = 48, Italy n = 90, Thailand n = 477, Latin America n = 157, Spain n = 426, Europe n = 364, Uganda n = 249, South America n = 210, Tanzania
9.8 yrs, 4.7 yrs 12 yrs, 5.1 yrs 10.6 yrs, 3yrs 6.8 yrs, 2.4 yrs 13 yrs, 10.2 yrs 12.2 yrs, 5.2 yrs 8 yrs, 3.8 yrs 7.1 yrs (mean), 5.6 yrs 9.6 yrs, NR
26% (7.5, 8.8 and 6.4%) 33% (8, 6 and 19%) 65% (13, 29.9 and 22.1%) NR 40.5% (20.4, 5.1 and 14.6%) 42% (28 and 27%; MP NR) 27% (22 and 4.7%; MP NR) NR 30% (19, 3.8 and 7.1%)
21% 14.6% 12% 29.4% 24.8% 16.7% 28.6% 33.6% NR
27% 16.7% 11% 20.5% 23.9% 12.4% 5.8% 12.7% NR
Prevalence of Insulin Resistance NR 52% 4% NR 20% 1.6% 0% 6.8% NR
LA indicates lipoatrophy; LH, lipohypertrophy; MP, mixed pattern; TC, total cholesterol; NR, not reported.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pidj.com | S37
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
Fortuny et al
TABLE 2. (a) Recommendations Regarding Healthy Lifestyle Habits; and (b) Monitoring Proposal for Metabolic and Renal Adverse Effects in the Routine Follow-up of HIV-infected Children and Adolescents (a) Calorically appropriate low-fat diet, especially in patients with high BMI or dyslipidemias Ensure sufficient calcium and vitamin D intake Regular aerobic exercise; weight-bearing exercise from puberty Smoking avoidance or cessation Avoidance of other toxic substances, illicit or not (b) Body weight and height at each visit, calculate BMI; blood pressure measurement at least every 3–6 months Clinical assessment of metabolic toxicities at each visit, including abnormal body fat distribution patterns In the absence of symptoms, lactate determination is not recommended Laboratory monitoring for dyslipidemias and glucose homeostasis: before and early after (4–12 weeks) HAART initiation or switch, and at least every 6–12 months thereafter if no ongoing toxicity concerns Periodic measurement of vitamin D, especially when several risk factors for low BMD coincide If available, DXA scans should be performed periodically, every 1–2 years, especially among children with abnormal findings and/or when several risk factors for low BMD coincide Urinalysis (proteinuria and glucosuria) and calculate estimated glomerular filtration rate at least every 6–12 months; every 3–6 months if tenofovir is used
protease inhibitor with either nevirapine or efavirenz, have shown disappointing results to date.38,41,42 Although the data are still insufficient, corrective surgery of severe facial lipoatrophy has been successfully used in teenagers43 and has become widely used in the adult patient, for central fat accumulation as well as facial.44 Other therapeutic possibilities should not be considered in children out of clinical trials (ie, recombinant growth hormone, testosterone, leptin, metformin and thiazolinediones).
DYSLIPIDEMIAS Normal lipid values for children are defined according to percentile cutpoints within population distributions. It should be kept in mind that age, sex and race have significant effects on lipid levels, raising doubts about the utility of fixed screening cutpoints. Twelve-hour fasting lipid level cutpoints commonly used for HIV infection in the pediatric age are as follows: triglyceride (TG) >100 mg/dL up to 10 years of age and >130 mg/dL in older patients, total cholesterol >200 mg/dL, low-density lipoprotein >130 mg/dL and high-density lipoprotein ≤40 mg/dL.45 Lipid abnormalities were described early in the HIVinfected pediatric population and have been associated mainly with exposure to protease inhibitors (especially with ritonavir boosting), stavudine and efavirenz, fat maldistribution and HIV disease,46–48 even in very young infants.15 HIV-unrelated factors linked to dyslipidemias have also been identified in this population, such as family history, inadequate diet, sedentarism, obesity, hypertension and smoking. Dyslipidemias may develop a few weeks after beginning therapy and are not associated with symptoms, although they will probably play a central role in future atherosclerotic disease. Again, wide ranges in prevalence of dyslipidemias have been reported in HIV-infected pediatric patients. In 2005, the PACTG 219C Study Group reported a 13% prevalence of hyperTC in the largest cohort of HIV-infected children and adolescents analyzed to date;46 later on, the same group described a resolution to normal total cholesterol within 2 years of follow-up in one-third of their patients.49 Other study figures for hyperTC and hyperTG range from 5.8% to 27% and from 12% to 33.6%, respectively (Table 1). Adolescents living with HIV infection should undergo regular overnight fasting lipid panel monitoring when initiating or changing HAART and at least every 6–12 months thereafter, together with routine blood analysis. In younger children, regular monitoring can be based on nonfasting samples; however, it should be kept in mind that nonfasting falsely elevated TG and low-density lipoprotein levels can be observed.
S38 | www.pidj.com
The same recommendations on lifestyle habits apply to the prevention of dyslipidemia (Table 2a), together with assessment of additional cardiovascular risk factors. Besides these, children with elevated TG levels should follow a diet with low simple carbohydrates. Some studies in children have reported improvements in lipid results upon switching from a protease inhibitor to a nonnucleoside.41,50–52 When possible, lipid-friendly ARVs should be chosen for treatment; in this respect, darunavir and atazanavir, even when ritonavir-boosted, as well as new ARV drug classes, have shown better lipid effects than older protease inhibitors in adult studies. When dyslipidemia occurs, a proper 3- to 6-month trial period of lifestyle modifications is recommended before any change in HAART or pharmacologic treatment is implemented. Any new treatment risk must be balanced against potential benefits (pharmacokinetic interactions, adverse events, pill burden, etc.). Lipid-lowering therapies should always be managed together with a lipid specialist and according to specific guidelines.10,11,35 Statins (pravastatin, atorvastatin, fluvastatin and rosuvastatin) are used to lower cholesterol levels by inhibiting the enzyme HMG-CoA reductase and some of them are approved for children aged 10 years or more. Side effects of statins include muscle pain, and increased risk of diabetes and hepatitis. All of them but pravastatin show multiple drug interactions with ARV. Ezetimibe acts by decreasing cholesterol absorption in the small intestine, and it is used alone or as add-on therapy with statins; it is also available for adolescents. Fibrates and omega-3 fatty acids are used in adults as secondary therapies of hyperTG but are not approved for the pediatric age. To date, the use of these drugs in the setting of pediatric HIV infection has been very scarcely reported.49,53
INSULIN RESISTANCE In contrast to adults, in children and youth living with HIV the spectrum of insulin resistance, asymptomatic hyperglycemia and diabetes mellitus appears to be a relatively uncommon finding. Insulin resistance is defined by means of elevated insulin levels or with the homeostatic model of insulin resistance (HOMA-IR = fasting insulin (μIU/mL) × fasting glucose (mg/dL)/405); cutoffs of 2.5 and 4.0 for HOMA-IR values are used in Tanner 1 and Tanner 2 and higher subjects, respectively. The prevalence of insulin resistance in different cohorts ranges from 0% to 52%, but elevated fasting hyperglycemia (>100 mg/dL) has only rarely been reported (Table 1).8,54–58 Several risk factors for the development of insulin resistance have been identified, with ARV use (mainly protease inhibitors and thymidine analogues) and elevated body mass index © 2015 Wolters Kluwer Health, Inc. All rights reserved.
The Pediatric Infectious Disease Journal • Volume 34, Number 5, May 2015
being the most important, but also fat maldistribution, advanced HIV disease, hepatitis C virus coinfection, family history, dyslipidemias, older age and puberty.2,54–56,59 To the best of our knowledge, no cases of frank diabetes mellitus type 2 have been reported in HIV-infected patients in the pediatric age, but typical symptoms to bear in mind are polydipsia, polyuria, polyphagia and changes in body habitus. Acanthosis nigricans, a nonsymptomatic skin condition due to hyperinsulinemia that causes one or more areas of skin to darken, thicken and feel velvety, seldom occurs in this population.60 In our experience, complete resolution of A. nigricans is observed upon protease inhibitor cessation. In the follow-up of impaired glucose metabolism in HIVinfected pediatric patients, only random or overnight fasting plasma glucose levels are recommended at HAART initiation or switch and every 6–12 months thereafter. Whenever random/fasting glucose levels over 140/100 mg/dL are observed or symptoms arise, referral to a pediatric endocrinologist is advisable. Determination of insulin levels or glycated hemoglobin levels, and glucose tolerance tests, are not routinely recommended. Primary prevention and initial treatment of glucose metabolism disturbances include lifestyle modifications (Table 2a) and avoidance or switching of certain ARV drugs (stavudine, indinavir, etc). An observational switch study in HIV-infected children showed improvements in glucose homeostasis parameters following protease inhibitor replacement.61 The use of metformin (licensed in children aged 10 years or more) and thiazolindinediones (rosiglitazone and pioglitazone; not approved below the age of 18 years) has not been reported in HIV-infected children to date. The former should be used with caution, because of the risk of lactic acidosis.
CARDIOVASCULAR AND CEREBROVASCULAR RISK HIV-infected adult patients are at a higher risk of early cardiovascular and cerebrovascular disease than the general population, due to HIV infection itself and also to ARV use.62,63 In recent years, there has been growing evidence of an increased global risk of premature cardiovascular disease also in children and adolescents with vertically acquired HIV infection. The overweight/obesity epidemic due to lifestyle factors among youth is also affecting the HIV-infected pediatric population, with prevalence rates up to 8.4% in Spain5 and 12% in the United States.64 Together with obesity and other HIV-unrelated factors, abnormal body fat distribution, lipid abnormalities and altered glucose metabolism lead to increased risk of future cardiovascular disease in children.65 For instance, excess trunk fat in healthy children was associated with elevated blood pressure levels, lipid levels and insulin levels in the Bogalusa Heart Study.66,67 Finally, uncontrolled HIV infectionrelated inflammatory and immunologic factors also contribute to cardiovascular risk.68 Recently, by means of the Pathobiological Determinants of Atherosclerosis in Youth (PDAY) scoring system,69 Patel et al70 described an increased coronary/abdominal aorta disease risk in 48%/24% of a large cohort of HIV-infected adolescents in the United States, with the highest scores being predicted by HIV disease severity and boosted protease inhibitor use. The cardiovascular toxicity associated with abacavir use remains controversial. In line with the former, several authors have reported increased carotid intima-media thickness and arterial stiffness as assessed by pulse wave velocity among HIV-infected children and adolescents when compared with healthy controls, often together with elevated metabolic and inflammatory markers of atherosclerotic disease, such as high-sensitive C-reactive protein, lipid levels, glycated hemoglobin and activated T cells.71–74 Miller et al75 also reported differences in markers of coagulant and endothelial dysfunction. Finally, more © 2015 Wolters Kluwer Health, Inc. All rights reserved.
ARV-related Comorbidities in HIV+ Children
than 50% of a small cohort of mostly vertically HIV-infected young adults showed coronary artery abnormalities consisting of luminal narrowing and irregularity of the coronary vessel wall with cardiac magnetic resonance techniques.76 Fortunately, cerebrovascular and cardiovascular clinical events in perinatally HIV-infected patients are still rare, although an increase in incidence is foreseeable in the coming decades.
LOW BMD AND OSTEOPOROSIS Low BMD is a recognized metabolic complication of HIV disease that has been noted more frequently in the era of HAART.77 A meta-analysis demonstrated a 15% prevalence of osteoporosis and a 67% prevalence of low BMD in HIV-infected adults some years ago;77 recently, a modest increase in incident fractures was also associated with HIV infection.78 Although low BMD has also been described in naïve patients, exposure to protease inhibitor-based HAART regimens and especially to tenofovir was associated with higher rates of osteoporosis when compared with other drugs.79,80 During childhood and adolescence, bone mass increases extensively through longitudinal growth. Several factors influence the absolute amount of peak bone mass attained, including gender, ethnicity, genetics, weight-bearing physical activity, dietary intake of calcium and vitamin D and pubertal development through hormonal influence. Peak bone mass occurs around late adolescence and, when suboptimal, becomes an important predictor of the future risk of osteoporosis and fractures in adulthood. DXA is the preferred method for assessing BMD in children and adolescents, and the spine and total body are the preferred skeletal sites. DXA values in children should always be adjusted using Z score height, especially in patients with growth delay, such as some HIV-infected patients, and they must be compared with an appropriate reference data set. In the pediatric age, the diagnosis of osteoporosis should not be made on the basis of densitometric criteria alone; the presence of vertebral compression (crush) fractures or low-energy peripheral long bones fractures is mandatory, together with a BMD Z score equal or below -2.0 (http://www.iscd. org/2013-iscd-official-positions-pediatric/). A BMD Z score equal to or below -2.0 alone defines low BMD in children; the term osteopenia should not be used. In vertically infected HIV-positive children and adolescents, several conditions associated with decreased BMD may coexist together with HIV infection, such as lack of mobility, prematurity, delayed puberty and growth and low body mass. Low BMD values have been consistently reported among different HIV-infected pediatric cohorts when compared with HIV-uninfected peers throughout the world, with prevalence rates ranging from 7% to 32%.81–84 In these studies, time on HAART, use of tenofovir, stavudine and protease inhibitors, advanced HIV disease and late pubertal stages were identified as risk factors for low BMD,81–85 whereas vitamin D status was not.81–83 Importantly, low BMD was also described among prepubertal patients,84 calling attention to the high risk of sub-optimal peak BMD should this situation remain unchanged through puberty. Vitamin D plays a central role in bone growth and remodeling. Highly variable rates of hypovitaminosis D (
