Diabetes mellitus in HIV-infected patients: Fasting glucose, A1c, or oral glucose tolerance test - which method to choose for the diagnosis?

RESEARCH ARTICLE Open Access Diabe es melli us in HIV-in ec ed pa ien s: as ing glucose, A1c, o o al glucose ole ance es –which me hod o choose o he diagnosis? Ana Ri a Coelho 1* , Flá ia And eia Mo ei a 1 , Ana C is ina San os 2,3 , And é Sil a-Pin o 4,5 , An ónio Sa men o 4,5 , Da ide Ca alho 6 and Paula F ei as 6 Abs ac Backg ound: An i e o i al he apy d ama ically educed HIV- ela ed mo bidi y and mo ali y, p olonging he li espan o HIV-in ec ed pa ien s. G ea e du a ion o in ec ion and exposu e o an i e o i al he apy makes hese pa ien s suscep ible o adi ional ca dio-me abolic isk ac o s and pa hologies. The op imal diagnos ic p o ocol o Diabe es Melli us in hese pa ien s is s ill con o e sial. Haemoglobin A1c (HbA1c) has been shown o unde es ima e glycaemia le els and he o al glucose ole ance es (OGTT) has been shown o e eal cases o glucose me abolism dis u bances in pa ien s wi h no mal as ing glucose. Thus, his s udy aimed o de e mine he p e alence o p ediabe es and diabe es in a popula ion o HIV-in ec ed pa ien s unde going combined an i e o i al he apy, using h ee di e en diagnos ic me hods ( as ing glucose, OGTT and HbA1c), o de e mine he ag eemen be ween he di e en me hods and he cha ac e is ics associa ed wi h each one. Me hods: This s udy analyzed 220 HIV-in ec ed pa ien s on an i e o i al he apy. Pa ien cha ac e is ics we e collec ed using a s anda dized p o ocol. Dis u bances o glucose homeos asis we e de ined by he ADA 2017 c i e ia. Pa ien s we e cha ac e ized acco ding o he p esence o absence o clinical lipodys ophy, and dis ibu ed in o ou di e en ca ego ies, acco ding o he p esence, o absence o ei he clinical lipoa ophy, o abdominal p ominence. Insulin esis ance was assessed by HOMA-IR and QUICKI indexes. Ag eemen be ween he diagnos ic me hods was assessed by Cohen’s kappa coe icien . Resul s: The e we e no pa ien s diagnosed wi h diabe es wi h HbA1c. 5.9% p e alence was ob ained when OGTT was used, and 3.2% p e alence when as ing glucose was used. P ediabe es had a p e alence o 14.1% when using HbA1c, 24.1% when using OGTT, and 20% when using as ing glucose. In all h ee me hods, glucose homeos asis dis u bances we e associa ed wi h olde age and highe esis ance o insulin. Rega ding o he cha ac e is ics, associa ions a ied be ween he h ee me hods. The ag eemen be ween hem was ai , o sligh . Conclusions: We obse ed ha HbA1c was he me hod ha diagnosed he leas amoun o cases and ha OGTT was he one ha diagnosed he mos cases. Acco dingly, ou esul s indica e ha HbA1c unde es ima ed glycaemia le els in his popula ion and ha he use o OGTT migh allow an ea lie diagnosis o glucose homeos asis dis u bances, po en ially making i possible o a oid se e e complica ions o DM. Keywo ds: HIV in ec ion, Diabe es melli us diagnosis, Fas ing glucose, O al glucose ole ance es , HbA1c * Co espondence: [email p o ec ed] 1 Medical S uden . Facul y o Medicine, Uni e si y o Po o. Alameda P o . He nâni Mon ei o, 4200-319 Po o, Po ugal Full lis o au ho in o ma ion is a ailable a he end o he a icle © The Au ho (s). 2018 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Coelho e al. BMC In ec ious Diseases (2018) 18:309 h ps://doi.o g/10.1186/s12879-018-3221-7 Backg ound An i e o i al he apy (ART) d ama ically educes HIV- ela ed mo bidi y and mo ali y and p olongs he li espan o his in ec ed popula ion [1]. The g ea e su - i al o pa ien s li ing wi h he in ec ion makes hem mo e suscep ible o exposu e o he same adi ional ca dio-me abolic isk ac o s and pa hologies as he gen- e al popula ion. These complica ions a e likely no only co ela ed wi h age, bu also wi h he cumula i e expos- u e o ART [1–5]. Ca dio-me abolic pa hology is, in ac , becoming an inc easing p oblem associa ed wi h HIV in ec ion unde ART [2,5,6]. In a la ge-scale HIV popula ion s udy, De Wi e al. ound ha he incidence o new-onse diabe es melli us (DM) inc eased wi h cumula i e exposu e o ART [5]. In ac , i has been shown ha HIV popula ions can ha e up o a wo- old highe isk o DM when com- pa ed o he gene al popula ion [4]. Besides ART, ad- i ional isk ac o s ha e also been associa ed wi h he de elopmen o DM in his popula ion, such as amily his o y, obesi y, olde age, ace, abdominal p ominence (e.g. wais ci cum e ence) and s a in use [4,7,8]. In he gene al popula ion, DM has been iden i ied as a high- isk, and e y high- isk ac o o he de elopmen o ca dio ascula disease [9]. This was also demons a ed in he HIV popula ion, whe e i was e ealed ha DM is a isk ac o ha subs an ially inc eases he chance o de eloping co ona y hea disease, especially in cases o p olonged in ec ion [6]. Thus, i is essen ial o egula ly sc een o his condi ion in his popula ion, in o de o p e en hese kinds o complica ions. The op imal diagnos ic me hodology o DM in HIV pa ien s is s ill a con o e sial opic. The cu en Eu opean AIDS Clinical Socie y guidelines ecommend he e alua ion o as ing glucose as an ini ial assessmen o glucose me abolism in pa ien s wi h new HIV diag- nos ic and also p io o s a ing ART. An o al glucose ole ance es (OGTT), o haemoglobin A1c (HbA1c) measu emen is only ecommended i as ing glucose le els e eal p ediabe es [10]. Acco dingly, his s udy aims o de e mine he p e a- lence o p ediabe es and DM using h ee di e ences diagnos ic me hods ( as ing glucose, OGTT and HbA1c) in a popula ion o HIV-in ec ed pa ien s unde com- bined an i e o i al he apy. Addi ionally, i is ou aim o in es iga e which cha ac e is ics di e in each diag- nos ic me hod be ween he di e en g oups, and wha is he ag eemen be ween he di e en me hods. Me hods Pa icipan s As pa o a c oss-sec ional s udy, be ween 2005 and 2016, 220 non-ins i u ionalized HIV-in ec ed adul s, consecu i ely e e ed om he In ec ious Diseases Depa men , we e e alua ed a he Endoc inology Ou pa ien Clinic o São João Hospi al. Pa ien s we e in- cluded in he s udy on hei i s isi , and only pa ien s on combined an i e o i al he apy we e included. A his- o y o p e iously-diagnosed DM and use o an i-diabe ic he apy excluded pa ien s om ou s udy. The E hics Commi ee o Heal h o Hospi al São João app o ed his s udy and each pa ien p o ided w i en in o med consen . Clinical assessmen Fo each pa ien he ollowing in o ma ion was collec ed using a s anda dized p o ocol: age, known du a ion o HIV in ec ion and o combined an i e o i al he apy ex- posu e, cu en ype o ART, HIV in ec ion isk ac o s and cha ac e iza ion o he in ec ion, smoking his o y (pas , cu en , o ne e ), his o y o diabe es and hype - ension, and use o an i-diabe ic, an i-hype ensi e, and lipid lowe ing d ugs. We used he Cen e s o Disease Con ol and P e en ion (CDC) c i e ia o classi ying he deg ee o in ec ion [11]. Weigh , heigh and wais ci cum e ence we e mea- su ed, and Body Mass Index (BMI) was also calcula ed. Body weigh was measu ed using TANITA (Tani a®, model TBF 300), and scale and heigh was measu ed o he nea es cen ime e in he s anding posi ion using a wall s adiome e (Hol ain Limi ed C ymych, Dy ed®). BMI was calcula ed as weigh di ided by heigh squa ed (kg/m2). The wais ci cum e ence was measu ed midway be ween he lowes ib and iliac c es , a he end o a gen le expi a ion, wi h he pa ien s anding up igh , ace di ec ed o wa d and shoulde s elaxed. Clinical lipodys ophy was de ined as pe iphe al lipoa- ophy wi h o wi hou cen al a accumula ion assessed by bo h pa ien and p ac i ione [12]. Pa ien s wi h a leas one ligh , mode a e, o se e e subjec i e lipoa- ophic ea u e (iden i ied by lipoa ophy-speci ic phys- ical examina ion) we e asked o epo whe he he/she had any change in a in hei cheeks, he side o hei ace, legs, a ms, o bu ocks. Pa ien s we e classi ied as being wi hou pe iphe al lipoa ophy when none o he p e iously desc ibed ea u es we e p esen [13]. P esence o cen al a accumula ion o abdominal p ominence was de ined by he measu emen o wais ci cum e ence using he In e na ional Diabe es Fede a ion c i e ia o me abolic synd ome (wais ci cum e ence ≥94 cm o Eu opid men and ≥80 cm o Eu opid women). Pa ien s we e classi ied in o ou di e en ca ego ies, acco ding o he p esence o absence o ei he clinical lipoa ophy o abdominal p ominence: 1) no lipodys ophy –pa- ien s wi hou lipoa ophy and wi hou abdominal p om- inence; 2) isola ed cen al a accumula ion –pa ien s wi hou lipoa ophy and wi h abdominal p ominence; 3) lipoa ophy –pa ien s wi h lipoa ophy and wi hou Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 2 o 13 abdominal p ominence; 4) mixed o ms o lipodys ophy –pa ien s wi h lipoa ophy and wi h abdominal p omin- ence [13]. The clinical assessmen was pe o med by he same p ac i ione (PF). Labo a o y analysis A enous blood sample was aken a e a 12-h o e nigh as . The 12-h o e nigh as was con i med wi h he pa - icipan s p io o he collec ion o he blood sample. I he 12-h as ing pe iod was no obse ed he blood sample col- lec ion was escheduled. All he samples we e analyzed a he cen al labo a o y o ou hospi al. The measu emen s o o al choles e ol, low-densi y lipop o ein choles e ol, high-densi y lipop o ein choles e ol, iglyce ides, plasma glucose and HbA1c se um le els we e de e mined using comme cial ki s. Hepa i is C was diagnosed by se ological assays ha de ec an ibody o hepa i is C i us (an i-HCV). All pa ien s wi hou a p e ious diagnosis o diabe es we e submi ed o an OGTT. The OGTT was pe o med as desc ibed by he Wo ld Heal h O ganiza ion, using a glucose load con aining he equi alen o 75 g anhyd ous glucose dissol ed in wa e . The CD4 cell coun was de e mined by low cy ome y and plasma HIV-1 RNA loads we e measu ed by a quan- i a i e e e se ansc ip ase polyme ase chain eac ion (Roche Diagnos ic Sys ems, Inc., B anchbu g, NJ, USA), which has a lowe limi o de ec ion o 50 copies/mL. C i e ia o he de ini ion o dis u bances o glucose homeos asis Dis u bances o glucose homeos asis we e de ined by he Ame ican Diabe es Associa ion 2017 c i e ia [14]. Pa ien s we e di ided in o h ee g oups: no diabe es (No DM), p ediabe es and diabe es (DM). No DM was de- ined as as ing glucose <100 mg/dL, HbA1c <5.7% o 120 min plasma glucose <140 mg/dL du ing he OGTT. P ediabe es was de ined as as ing glucose be ween 100 and 126 mg/dL, HbA1c be ween 5.7 and 6.5% o 120 min plasma glucose be ween 140 and 200 mg/dL du ing he OGTT. DM was de ined as as ing glucose ≥126 mg/dL, HbA1c ≥6.5% o 120 min plasma glucose ≥200 mg/dL du ing he OGTT. Measu emen s o insulin esis ance Insulin esis ance was de ined by he homeos asis model assessmen o insulin esis ance (HOMA), and insulin sensi i i y by he quan i a i e insulin sensi i i y check index (QUICKI). These indexes we e calcula ed by he ollowing o mulas: HOMA-IR index = ( as ing plasma insulin × as ing plasma glucose)/22.5 [15] and QUICKI = 1/[log ( as ing insulin in mU/l) + log ( as ing plasma glucose in mg/dL)] [16]. Glucose was exp essed in mmol/L and insulin in μUI/mL. Insulin esis ance was de ined when he alue o HOMA >4 [8]. S a is ical analysis Quan i a i e a iables we e desc ibed as mean and s anda d de ia ion (SD), o median and in e qua ile ange (IQR), and we e compa ed using S uden - and ANOVA o Mann–Whi ney and K uskal-Wallis es s, as app op ia e. Ca ego ical a iables we e desc ibed as coun s and p opo ions, and compa ed using he chi-squa e o Fishe ’s exac es . The kappa coe icien was compu ed o analyze s a is ical ag eemen be ween he h ee di e - en diagnos ic me hods used o de ining p ediabe es and DM. S a is ical analysis was pe o med using SPSS e sion 24.0 so wa e (SPSS Inc., Chicago, Illinois, USA). All p obabili ies we e wo ailed, and p alues o <0.05 we e ega ded as signi ican . Resul s Baseline cha ac e is ics A o al o 220 HIV-in ec ed pa ien s unde ART we e e alua ed. The mean age o pa ien s included was 45.8 ± 11.5 yea s, and 60.5% o hem we e males. All he demo- g aphic and clinical cha ac e is ics accessed in his s udy a e p esen ed in Table 1, acco ding o he p esence o absence o clinical lipodys ophy. Pa ien s wi h clinical lipodys ophy we e olde [47.5 (11.29) s 43.82 (11.49) yea s; P= 0.017], had longe du a ion o he HIV in ec ion [9.0 (5.0) s 6.0 (6.0) yea s; p= 0.001] and o ART use [8.0 (5.0) s 5.0 (5.5) yea s; p< 0.001]. Rega ding an h opome ic measu es, pa ien s wi h clinical lipodys ophy had lowe weigh (p< 0.001), BMI ((p< 0.001), and wais ci cum e ence mean alues (p< 0.001). Hype ension was mo e e- quen in pa ien s wi h clinical lipodys ophy [45 (39.1) s 23 (21.9) %; p= 0.009], as we e cu en smoke s [56 (49.1) s 36 (34.3) %; p= 0.032], and iglyce ides me- dian alues we e signi ican ly highe han hose o pa- ien s wi hou clinical lipodys ophy [215.5 (214.5) s 171 (154) mg/dL; p= 0.007]. Wi h ega ds o pha ma- ceu ical he apy, he use o s a ins (p= 0.005) and ib a es (p= 0.054) was mo e equen in pa ien s wi h clinical lipodys ophy. No di e ences we e ound be ween pa ien s wi h o wi hou clinical lipodys ophy in e ms o gende , CD4+ cell coun , pe cen age o i al supp ession, p e alence o co-in ec ion wi h Hepa i is C, ype o isk ac o o he HIV ansmission, CDC clinical ca ego ies, ype o ART used, and lipid p o ile. Hemoglobin A1c No pa ien s we e diagnosed wi h DM using his me hod (HbA1c ≥6.5%). The e o e, in Table 2, he esul s e- ga ding wo g oups: no DM (HbA1c ≤5.7%) and p edia- be es (5.7 < HbA1c < 6.5) a e p esen ed. Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 3 o 13 In ou popula ion, 31 pa ien s (14.1%) we e diagnosed wi h p ediabe es. These pa ien s we e olde [50.35 (11.89) s 44.99 (11.30) yea s; p=0.016], and had a highe BMI [26.90 (4.46) s 25.01 (4.49) kg/m2; p= 0.031], compa ed o he no DM pa ien s, bu had lowe me- dian le el o iglyce ides [139 (133.8) s 203 (198.0) mg/dL, p= 0.010], and less equen ly used ib a e [4(12.9) s 61 (32.3)%; p= 0.048]. The alues ob ained Table 1 Sample’s baseline cha ac e is ics, acco ding o he p esence o Clinical Lipodys ophy (CL) Wi h CL Wi hou CL P alue n(%) 115 (52.3) 105 (47.7) Sex [n(%)] 0.099 Male 76 (66.1) 57 (54.3) Female 39 (33.9) 48 (45.7) Age [yea s, mean (SD)] 47.5 (11.3) 43.8 (11.5) 0.017 Du a ion o HIV in ec ion [yea s, median (IR)] 9 (5) 6 (6) 0.001 cART [yea s, median (IR)] 8 (5) 5 (5.5) < 0.001 Weigh [Kg, mean (SD)] 64.1 (12.8) 73.9 (12.6) < 0.001 Heigh [m, mean (SD)] 1.7 (0.1) 1.7 (0.1) 0.765 BMI [(kg/m2), mean (SD)] 23.6 (3.8) 27.2 (4.5) < 0.001 Wais ci cum e ence [cm, mean (SD)] 88.4 (10.8) 95.1 (12.1) < 0.001 CD4 cell coun [cells/mm3, median (IR)] 554 (385) 479 (300) 0.238 HIV RNA (<50) [n(%)] 100 (100) 92 (100) Hepa i is C co-in ec ion [n(%)] 34 (29.8) 30 (29.4) 0.999 Hype ension [n(%)] 45 (39.1) 23 (21.9) 0.009 HIV isk ac o [n(%)] 0.162 In a enous d ug use 1 (25) 3 (15.8) Homosexual con ac 0 (0) 2 (10.5) He e osexual con ac 2 (50) 14 (73.7) O he s 1 (25) 0 (0) CDC clinical ca ego ies [n(%)] 0.389 A 63 (54.8) 56 (53.3) B 1 (0.9) 4 (3.8) C 51 (44.3) 45 (42.9) ART [n(%)] IP 61 (53) 62 (59) 0.447 NNRTI 55 (47.8) 47 (44.8) 0.749 NRTI 113 (98.3) 97 (92.4) 0.051 Smoking his o y [n(%)] 0.032 Ne e 38 (33.3) 53 (50.5) Cu en 56 (49.1) 36 (34.3) Fo me 20 (17.5) 16 (15.2) To al choles e ol [mg/dL, mean (SD)] 221.9 (53.2) 227.8 (57.8) 0.433 LDL- choles e ol [mg/dL, mean (SD)] 129.4 (48.1) 140 (45.4) 0.097 HDL- choles e ol [mg/dL, mean (SD)] 46.5 (14.8) 49.5 (13.3) 0.123 T iglyce ides [mg/dL, median (IR)] 215.5 (214.5) 171 (154) 0.007 S a in use [n(%)] 33 (28.7) 13 (12.4) 0.005 Fib a e use [n(%)] 41 (35.7) 24 (22.9) 0.054 CL clinical lipodys ophy, cART combina ion an i e o i al he apy, BMI body mass index, ART an i e o i al he apy, PI p o ease inhibi o , NNRTI non-nucleoside e e se ansc ip ase inhibi o , NRTI nucleoside e e se ansc ip ase inhibi o , HOMA homeos a ic model assessmen , QUICKI quan i a i e insulin sensi i i y check index, SD s anda d de ia ion, IR in e qua ile ange Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 4 o 13 o HOMA-IR we e highe among pa ien s wi h p edia- be es, and he di e ence be ween he g oups was s a is- ically signi ican [2.51 (5.13) s 1.62 (1.46); p= 0.023]. The e we e no di e ences be ween he diagnos ic g oups wi h ega ds o sex, du a ion o HIV in ec ion o ART use, p esence o absence o clinical lipodys ophy, body composi ion ypes, wais ci cum e ence, CD4 cell coun , pe cen age o i al supp ession, p e alence o hepa i is C coin ec ion, CDC clinical ca ego y, ype o ART used, lipid p o ile, and equency o use o s a ins. Table 2 Sample’s cha ac e is ics, acco ding o he p esence o no DM o p ediabe es, accessed by HbA1c HbA1c P alue <5.7% 5.7–6.4% n(%) 189 (85.9) 31 (14.1) Sex [n(%)] 0.623 Male 116 (61.4) 17 (54.8) Female 73 (38.6) 14 (45.2) Age [yea s, mean (SD)] 45 (11.3) 50.4 (11.9) 0.016 Du a ion o HIV in ec ion [yea s, median (IR)] 8 (6) 8 (7.0) 0.698 cART [yea s, median (IR)] 6 (6.5) 7 (7.0) 0.136 Clinical lipodys ophy [n(%)] 0.909 Wi hou CL 91 (48.1) 14 (45.2) Wi h CL 98 (51.9) 17 (54.8) Body Composi ion [n(%)] 0.469 No lipodys ophy 28 (15.3) 3 (10) Isola ed cen al a accumula ion 59 (32.2) 10 (33.3) Lipoa ophy 52 (28.4) 6 (20) Mixed o m o lipodys ophy 44 (24) 11 (36.7) BMI [(kg/m2), mean (SD)] 25 (4.5) 26.9 (4.5) 0.031 Wais ci cum e ence [cm, mean (SD)] 90.9 (11.6) 95.2 (13.2) 0.065 CD4 cell coun [cells/mm3, median (IR)] 500 (345) 528 (312) 0.819 HIV RNA (<50) [n(%)] 165 (100) 27 (100) Hepa i is C co-in ec ion [n(%)] 57 (30.8) 7 (22.6) 0.474 CDC clinical ca ego ies [n(%)] 0.093 A 107 (56.6) 12 (38.7) B 5 (2.6) 0 (0) C 77 (40.7) 19 (61.3) ART [n(%)] IP 106 (56.1) 17 (54.8) 0.999 NNRTI 88 (46.6) 14 (45.2) 0.999 NRTI 182 (96.3) 28 (90.3) 0.152 HOMA-IR index [median (IR)] 1.6 (1.5) 2.5 (5.1) 0.023 QUICKI index [median (IR)] 0.4 (0.1) 0.3 (0.1) 0.023 To al choles e ol [mg/dL, mean (SD)] 225.6 (55.4) 219 (55.8) 0.544 LDL- choles e ol [mg/dL, mean (SD)] 133.5 (47.5) 140.2 (44.6) 0.473 HDL- choles e ol [mg/dL, mean (SD)] 48.3 (14.4) 45.8 (11.8) 0.374 T iglyce ides [mg/dL, median (IR)] 203 (198) 139 (133.8) 0.010 S a in use [n(%)] 41 (21.7) 5 (16.1) 0.640 Fib a e use [n(%)] 61 (32.3) 4 (12.9) 0.048 DM diabe es melli us, HbA1c glyca ed haemoglobin, CL clinical lipodys ophy, cART combina ion an i e o i al he apy, BMI body mass index, ART an i e o i al he apy, PI p o ease inhibi o , NNRTI non-nucleoside e e se ansc ip ase inhibi o , NRTI nucleoside e e se ansc ip ase inhibi o , HOMA homeos a ic model assessmen index, QUICKI quan i a i e insulin sensi i i y check index, SD s anda d de ia ion, IR in e qua ile ange Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 5 o 13 O al glucose ole ance es Wi h ega ds o OGTT, 53 pa ien s (24.1%) we e diagnosed wi h p ediabe es (140 mg/dL < glucose a 120 min < 200 mg/dL), and 13 pa ien s (5.9%) wi h DM (glucose a 120 min ≥200 mg/dL). In Table 3, he esul s ega ding he h ee ca ego ies o glucose homeos asis a e p esen ed. Pa ien s diagnosed wi h DM we e olde han hose wi h p ediabe es, and hese, in u n, we e olde han hose wi hou DM [55.00 (13.55) s 50.36 (12.13) s 43.38 (10.24) yea s; p< 0.001]. A p og essi e s age o glucose me abolism diso de appea s ela ed o he use o p o ease inhibi o s [76.9% s 67.9% s 50%; p= 0.023], and he opposi e is obse ed wi h he use o non-nucleoside e e se ansc ip ase inhibi o s [23.1% s 32.1% s 53.2%; p= 0.005). The HOMA-IR index was highes among he p edia- be es g oup, and lowes among he no DM g oup [2.37 (2.49) s 1.94(3.17) s 1.41 (1.54); p< 0.001]. The e we e no di e ences be ween he diagnos ic g oups ega ding sex, du a ion o HIV in ec ion o ART use, p esence o absence o clinical lipodys ophy, body composi ion ypes, BMI, wais ci cum e ence, CD4 cell coun , pe cen age o i al supp ession, p e alence o hepa i is C coin ec ion, CDC clinical ca ego y, lipid p o- ile, and equency o use o s a ins o ib a es. Fas ing glucose Fo y ou pa ien s (20%) we e diagnosed wi h p ediabe- es (100 mg/dL < as ing glucose< 126 mg/dL), and se en pa ien s (3.2%) we e diagnosed wi h DM ( as ing glu- cose≥126 mg/dL). In Table 4, we p esen ou esul s e- ga ding he h ee ca ego ies o glucose homeos asis. We obse ed ha he e was a signi ican di e ence e- ga ding sex be ween he diagnos ic g oups, as all he pa- ien s diagnosed wi h DM we e men. In he p ediabe es g oup, 47.7% o he pa ien s we e woman. The di e - ences in age we e also s a is ically signi ican , wi h he oldes pa ien s being in he p ediabe es g oup, and he younges pa ien s being in he no DM g oup [no DM 43.00 (15) s p ediabe es 51.50 (15) s DM 45.00 (14) yea s; p= 0.027]. Rega ding he di e en ca ego ies o body composi ion, we obse e ha pa ien s wi hou DM diagnosis had he highes p opo ions o pa ien s in he ca ego ies “No lipodys ophy”(15.8%) and “Lipoa ophy” (31.5%), he p ediabe es g oup had he highes pe cen - age o pa ien s in he “Isola ed cen al a accumula ion” (53.7%) ca ego y, and he DM g oup had he highes pe cen age o pa ien s in he ca ego y “Mixed o m o lipodys ophy”(71.4%). Wais ci cum e ence was signi i- can ly lowe in he no DM g oup [noDM 88.0 (16.5) s p ediabe es 95.0 (13.5) s DM 95.0 (15.0); p= 0.005]. Pa ien s wi h hepa i is C coin ec ion we e mo e equen ly classi ied as no DM g oup (33.5%), and less equen ly in he p ediabe es g oup (14.3%). The HOMA-IR index was highes among he DM g oup and lowes among he no DM g oup [no DM 1.45 (1.31) s p ediabe es 3.18 (2.57) s DM 9.27 (6.79); p< 0.001]. The e we e no di e ences be ween he diagnos ic g oups ega ding du a ion o HIV in ec ion o ART use, p esence o absence o clinical lipodys ophy, BMI, CD4 cell coun , pe cen age o i al supp ession, CDC clinical ca ego y, ype o ART used, lipid p o ile, and equency o use o s a ins o ib a es. Ag eemen analyses Kappa coe icien s we e compu ed o es ima e he ag eemen be ween he h ee diagnos ic de ini ions. In Table 5we p esen he esul s om he analyses be ween OGTT and HbA1c. The kappa coe icien alue was 0.141 (p= 0.025), which co esponds o only a sligh ag eemen . In Table 6, esul s om he ag eemen be ween as ing glucose and HbA1c a e p esen ed, and he kappa coe icien was 0.013 (p= 0.848). In Table 7, we p esen he esul s om he analyses be ween OGTT and as ing glucose. In his pai he kappa alue was 0.206 (p< 0.001), which is conside ed o be a ai ag eemen . Discussion To ou knowledge, his is he i s s udy ha has been ca ied ou wi h HIV-in ec ed pa ien s, ha combines he use o as ing glucose, OGTT and HbA1c, o es ab- lish he diagnosis o glucose homeos asis dis u bances. Analyzing he p e alence o DM and p ediabe es h oughou hese di e en me hods, we obse e ha he esul s a ied conside ably. Rega ding he diagnosis o DM, no pa ien s we e iden i ied when HbA1c was used, 13 pa ien s (5.9% p e alence) when OGTT was used, and se en pa ien s (3.2% p e alence) when as ing glu- cose was used. P ediabe es had a p e alence o 14.1% (31 pa ien s) when using HbA1c, 24.1% (53 pa ien s) when using OGTT, and 20% (44 pa ien s) when using as ing glucose. Thus, we can conclude ha HbA1c was he me hod ha leas diagnosed cases, and ha OGTT was he one ha diagnosed he mos cases. In acco dance wi h ou indings, HbA1c has been ound o unde es ima e glycemic le els in HIV-in ec ed pa ien s when compa ed wi h o he ypes o diagnos ic me hods, in se e al s udies [4,17–22]. Possible explana- ions o he lowe han expec ed HbA1c alues in hese pa ien s ha e been hypo hesized. Fo example, low hemoglobin alues [19], and si ua ions ha sho en e y h ocy e li espan, such as hemolysis o some hemo- globinopa hies, ha e been associa ed wi h lowe HbA1c alues [18]. Diop e al. ha e ound ha he disco dance HbA1c- as ing glucose was posi i ely co ela ed wi h he mean cell olume, and ha hemolysis, diagnosed by a Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 6 o 13 e y low hap oglobin le el, had a highe p e alence in he HIV-in ec ed pa ien s [18]. In ac , his ela ion- ship be ween mean cell olume and HbA1c- as ing glucose disco dance has been obse ed by se e al au- ho s [17,18,21–23], and Glesby e al. alida ed i when hey obse ed ha highe mean cell olume alues eme ged as he single mos impo an ac o associa ed wi h a lowe HbA1c han p edic ed by as ing glucose. High mean cell olume, as a ma ke o a g ea e p opo - ion o younge e y h ocy es ha had a sho e ime o Table 3 Sample’s cha ac e is ics acco ding o he p esence o no DM, p ediabe es and DM, accessed by OGTT Glucose a 120 min P alue <140 mg/dL 140–200 mg/dL ≥200 mg/dL n(%) 154 (70) 53 (24.1) 13 (5.9) Sex [n(%)] 0.061 Male 101 (65.6) 26 (49.1) 6 (46.2) Female 53 (34.4) 27 (50.9) 7 (53.8) Age [yea s, mean (SD)] 43.4 (10.2) 50.4 (12.1) 55 (13.6) < 0.001 Du a ion o HIV in ec ion [yea s, median (IR)] 8 (6) 7 (6) 10 (8) 0.980 cART [yea s, [median (IR)] 6 (5.3) 6 (7) 9 (9) 0.566 Clinical lipodys ophy [n(%)] 0.148 Wi hou CL 78 (50.6) 24 (45.3) 3 (23.1) Wi h CL 76 (49.4) 29 (54.7) 10 (76.9) Body Composi ion [n(%)] 0.078 No lipodys ophy 23 (15.5) 7 (13.5) 1 (7.7) Isola ed cen al a accumula ion 50 (33.8) 17 (32.7) 2 (15.4) Lipoa ophy 45 (30.4) 11 (21.2) 2 (15.4) Mixed o m o lipodys ophy 30 (20.3) 17 (32.7) 8 (61.5) BMI [(kg/m2), [median (IR)] 24.5 (5.8) 25.4 (6.4) 25.5 (3.3) Wais ci cum e ence [cm, median (IR)] 90.5 (16) 91 (17.8) 95 (8.5) 0.687 CD4 cell coun [cells/mm3, [median (IR)] 512.5 (336) 500 (311) 456 (509) 0.569 HIV RNA (<50) [n(%)] 138 (100) 44 (100) 10 (100) Hepa i is C co-in ec ion [n(%)] 47 (30.9) 15 (28.8) 2 (16.7) 0.655 CDC clinical ca ego ies [n(%)] 0.252 A 90 (58.4) 24 (45.3) 5 (38.5) B 4 (2.6) 1 (1.9) 0 (0) C 60 (39) 28 (52.8) 8 (61.5) ART [n(%)] IP 77 (50) 36 (67.9) 10 (76.9) 0.023 NNRTI 82 (53.2) 17 (32.1) 3 (23.1) 0.005 NRTI 148 (96.1) 51 (96.2) 11 (84.6) 0.188 HOMA-IR index [median (IR)] 1.4 (1.5) 2.4 (2.5) 1.9 (3.2) < 0.001 QUICKI index [median (IR)] 0.4 (0.1) 0.3 (0.1) 0.4 (0.1) < 0.001 To al choles e ol [mg/dL, median (IR)] 224 (67) 224 (79) 234 (86) 0.743 LDL- choles e ol [mg/dL, median (IR)] 130 (70) 135 (66) 156 (79) 0.843 HDL- choles e ol [mg/dL, median (IR)] 49 (19) 44 (20) 52 (15) 0.314 T iglyce ides [mg/dL, median (IR)] 186 (176.5) 223 (258.5) 170 (101) 0.131 S a in use [n(%)] 31 (20.1) 9 (17) 6 (46.2) 0.072 Fib a e use [n(%)] 43 (27.9) 21 (39.6) 1 (7.7) 0.061 DM diabe es melli us, OGTT o al glucose ole ance es , CL clinical lipodys ophy, cART combina ion an i e o i al he apy, BMI body mass index, ART an i e o i al he apy, PI p o ease inhibi o , NNRTI non-nucleoside e e se ansc ip ase inhibi o , NRTI nucleoside e e se ansc ip ase inhibi o , HOMA homeos a ic model assessmen index, QUICKI quan i a i e insulin sensi i i y check index, SD s anda d de ia ion, IR in e qua ile ange Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 7 o 13 become glyca ed, sugges a g ea e ed blood cell u no e in he HIV-in ec ed pa ien s [17]. The e en ual ela ion- ship wi h d ugs used in he ea men o HIV in ec ion wi h hese hema ologic indings is di icul o s udy, as ART is gene ally used in combina ion. OGTT was he me hod ha diagnosed he highes pe cen age o pa ien s wi h DM and p ediabe es. Simila esul s ha e been ound in he li e a u e [24–26]. Giano i e al. demons a ed in hei s udy ha OGTT e ealed ha 11% o hei coho wi h long-s anding Table 4 Sample’s cha ac e is ics acco ding o he p esence o no DM, p ediabe es and DM, accessed by as ing glucose Fas ing glucose P alue <100 mg/dL 100–126 mg/dL ≥126 mg/dL n(%) 169 (76.8) 44 (20) 7 (3.2) Sex [n(%)] 0.049 Male 103 (60.9) 23 (52.3) 7 (100) Female 66 (39.1) 21 (47.7) 0 (0) Age [yea s, [median (IR)] 43 (15) 51.5 (15) 45 (14) 0.027 Du a ion o HIV in ec ion [yea s, [median (IR)] 8 (6) 8 (5.8) 8 (6) 0.782 cART [yea s, [median (IR)] 6 (7) 6 (5.8) 8 (6) 0.408 Clinical lipodys ophy [n(%)] 0.082 Wi hou CL 76 (45) 27 (61.4) 2 (28.6) Wi h CL 93 (55) 17 (38.6) 5 (71.4) Body Composi ion [n(%)] 0.004 No lipodys ophy 26 (15.8) 4 (9.8) 1 (14.3) Isola ed cen al a accumula ion 46 (27.9) 22 (53.7) 1 (14.3) Lipoa ophy 52 (31.5) 6 (14.6) 0 (0) Mixed o m o lipodys ophy 41 (24.8) 9 (22) 5 (71.4) BMI [(kg/m2), [median (IR)] 24.4 (5.8) 26.1 (5.5) 26 (3.4) 0.052 Wais ci cum e ence [cm, [median (IR)] 88 (16.5) 95 (13.5) 95 (15) 0.005 CD4 cell coun [cells/mm3, [median (IR)] 486 (344) 525 (298) 605 (382) 0.643 HIV RNA (<50) [n(%)] 151 (100) 35 (100) 6 (100) Hepa i is C co-in ec ion [n(%)] 56 (33.5) 6 (14.3) 2 (28.6) 0.034 CDC clinical ca ego ies [n(%)] 0.398 A 89 (52.7) 27 (61.4) 3 (42.9) B 3 (1.8) 2 (4.5) 0 (0) C 77 (45.6) 15 (34.1) 4 (57.1) ART [n(%)] IP 96 (56.8) 23 (52.3) 4 (57.1) 0.885 NNRTI 79 (46.7) 20 (45.5) 3 (42.9) 0.999 NRTI 163 (96.4) 41 (93.2) 6 (85.7) 0.157 HOMA-IR index [median (IR)] 1.5 (1.3) 3.2 (2.6) 9.3 (6,8) < 0.001 QUICKI index [median (IR)] 0.4 (0.1) 0.3 (0.04) 0.3 (0.04) < 0.001 To al choles e ol [mg/dL, median (IR)] 220 (66) 240.5 (86) 234 (79) 0.061 LDL- choles e ol [mg/dL, median (IR)] 127 (67) 149 (62) 146 (74) 0.136 HDL- choles e ol [mg/dL, median (IR)] 47 (18) 49 (22) 41 (22) 0.651 T iglyce ides [mg/dL, median (IR)] 189.5 (197.8) 214.5 (207.8) 184 (88) 0.975 S a in use [n(%)] 35 (20.7) 8 (18.2) 3 (42.9) 0.312 Fib a e use [n(%)] 50 (29.6) 15 (34.1) 0 (0) 0.200 DM diabe es melli us, CL clinical lipodys ophy, cART combina ion an i e o i al he apy, BMI body mass index, ART an i e o i al he apy, PI p o ease inhibi o , NNRTI non-nucleoside e e se ansc ip ase inhibi o , NRTI nucleoside e e se ansc ip ase inhibi o , HOMA homeos a ic model assessmen index, QUICKI quan i a i e insulin sensi i i y check index, SD s anda d de ia ion, IR in e qua ile ange Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 8 o 13 HIV in ec ion had p ediabe es o DM, undiagnosed on he basis o as ing glucose le els alone [24], while Seang e al. de ec ed a 31% ela i e inc ease in he p e alence o DM diagnosis among HIV-in ec ed women [25]. Epidemiological e idence has also suppo ed his obse - a ion in he gene al popula ion, especially among olde pa ien s [27]. In ligh o his knowledge, he 2017 Eu opean AIDS Clinical Socie y Guidelines ecommend ha HIV-in ec ed pa ien s wi h a as ing glucose diagno- sis o p ediabe es should ca y ou an OGTT in o de o iden i y o e diabe es [10]. Wi h ega ds o he cha ac e is ics ha di e ed signi i- can ly be ween he diagnos ic g oups, we obse ed ha he e we e some ha we e impo an ac oss all h ee me hods, while o he s a ied speci ically, depending on he me hod used o he assessmen . Sex was a ac o ha showed associa ion wi h he diag- nosis only when as ing glucose was used. Wi h his me hod, he o al numbe o pa ien s ha had he diag- nosis o DM we e male, and he e was also a highe p e alence o male (52.3% e sus 47.7% emale) in he p ediabe es g oup. Some au ho s ha e epo ed a lack o signi ican di e ences be ween sexes when conside ing he isk ac o s o glucose homeos asis dis u bances in HIV pa ien s [24], while o he s, in ag eemen wi h ou indings, s a ed ha male sex was associa ed wi h in- c eased isk o new-onse DM [5,28]. In all h ee me hods, a p og essi e s age o glucose homeos asis dis u bance was associa ed wi h olde age. This inding is cong uen wi h he g ea majo i y s udies done on he subjec [1–5,7,28,29]. This has been highligh ed in ecen li e a u e ha , a e he in oduc ion o ART (which has d ama ically educed HIV- ela ed mo ali y and mo bidi y, subs an ially in- c easing longe i y), HIV-in ec ed indi iduals ha e a po en- ial o de eloping me abolic complica ions, which is compa able o ha o he gene al popula ion [1]. In hese pa ien s, he impo ance o adi ional ca dio-me abolic isk ac o s should be emphasized, as hese a e likely o exe an equal in luence on HIV-in ec ed pa ien s as hey do in he gene al popula ion [1]. On he o he hand, he e ha e been s udies ha sugges ha he aging p ocess migh be p ema u e o accele a ed in hese pa ien s, lead- ing o he mani es a ion o me abolic complica ions ea lie in li e [2]. This highligh s he impo ance o closely moni- o ing o he de elopmen o ca dio-me abolic abno mal- i ies in hese pa ien s. Rega ding he du a ion o HIV in ec ion and ART use, we obse ed ha , ega dless o he diagnos ic me hod used, no signi ican di e ence be ween he g oups was ob- se ed. This con as s wi h indings in he li e a u e, ha sugges ha a highe p e alence o DM is associa ed wi h a highe du a ion o HIV in ec ion and ART use [1,3,30], bu , is in ag eemen wi h a s udy done by A aújo e al., in which also, no associa ion be ween he du a ion o in ec- ion and he de elopmen o glucose homeos asis dis u - bances was ound [31]. BMI was conside ed o be signi ican ly highe in p edia- be es pa ien s, compa ed o he no DM pa ien s, bu only when HbA1c was used o he assessmen . This inding is co obo a ed by se e al au ho s [3–5,7,28–30] who de- sc ibed a signi ican associa ion be ween a highe BMI and he p esence o dis u bances o glucose homeos asis in HIV-in ec ed pa ien s. Addi ional o he BMI, abdom- inal a accumula ion o unk obesi y has been iden i ied as a ac o p ima ily associa ed wi h he p e alence o dis- u bances o glucose homeos asis in HIV-in ec ed pa ien s [8,25,26,30–33], jus as desc ibed in he gene al popula- ion. In ou s udy, when as ing glucose was used o he assessmen o diabe es, he e was a signi ican associa ion be ween wais ci cum e ence and he diagnosis. This pa - ame e was highe in DM and p ediabe es pa ien s, when compa ed wi h he no DM ones. Fu he mo e, in his diagnos ic me hod, we ob ained signi ican di e ences e- ga ding he classes o body composi ion, wi h mos signi i- can di e ences obse ed in he isola ed cen al a accumula ion (highes pe cen age o pa ien s in he p edi- abe es g oup) and mixed o ms o lipodys ophy (highes pe cen age o pa ien s in he DM g oup), which mi o he e ec o hese highe wais ci cum e ence le els in hese diagnos ic g oups. In ac , i has been epo ed ha abdominal a accumula ion is a majo con ibu o o glu- cose me abolism dis u bances when compa ed o lipoa o- phy associa ed wi h he acqui ed lipodys ophy o HIV in ec ion [8]. Endoc ine ac i i y o adipose issue akes a cen al place in he pa hogenesis o me abolic diso de s Table 5 Analysis o he ag eemen be ween HbA1c and Glucose a 120 min du ing an OGTT Glucose a 120 min To al <140 mg/dL 140–200 mg/dL HbA1c <5.7% 140 42 182 5.7–6.4% 14 11 25 To al 154 53 207 Kappa Coe icien = 0.141 (p= 0.025) HbA1c glyca ed haemoglobin, OGTT o al glucose ole ance es Table 6 Analysis o he ag eemen be ween HbA1c and as ing glucose Fas ing glucose To al <100 mg/dL 100–126 mg/dL HbA1c <5.7% 148 39 187 5.7–6.4% 21 5 26 To al 169 44 213 Kappa Coe icien = 0.013 (p= 0.848) HbA1c glyca ed haemoglobin Coelho e al. BMC In ec ious Diseases (2018) 18:309 Page 9 o 13