Correlates of thymus size and changes during treatment of children with severe acute malnutrition: a cohort study

RESEARCH ARTICLE Open Access Co ela es o hymus size and changes du ing ea men o child en wi h se e e acu e malnu i ion: a coho s udy Ma en Johanne Heilsko Ry e 1* , Hani a Namusoke 2 , Ch is ian Ri z 1 , Kim F. Michaelsen 1 , And é B iend 1,3 , Hen ik F iis 1 and Do he Jeppesen 4 Abs ac Backg ound: The impai men o immune unc ions associa ed wi h malnu i ion may be one eason o he high mo ali y in child en wi h se e e acu e malnu i ion (SAM), and hymus a ophy has been p oposed as a ma ke o his immunode iciency. The aim o his s udy was o iden i y nu i ional and clinical co ela es o hymus size in child en wi h SAM, and p edic o s o change in hymus size wi h nu i ional ehabili a ion. Me hods: In an obse a ional s udy among child en aged 6–59 mon hs admi ed wi h SAM in Uganda, we measu ed hymus a ea by ul asound on hospi al admission o ea men wi h F75 and F100, on hospi al discha ge and a e 8 weeks o nu i ional ehabili a ion wi h eady- o-use he apeu ic ood, as well as in well-nou ished heal hy child en. We in es iga ed an h opome ic, clinical, biochemical and ea men - ela ed co ela es o a ea and g ow h o he hymus. Resul s: Eigh y- i e child en wi h SAM wi h a median age o 16.5 mon hs we e included. On admission 27% o he child en had a hymus unde ec able by ul asound. Median hymus a ea was 1.3 cm 2 in malnou ished child en, and 3.5 cm 2 in heal hy child en (p< 0.001). Mos an h opome ic z-sco es, hemoglobin and plasma phospha e co ela ed posi i ely wi h hymus a ea. Thymus a ea co ela ed nega i ely wi h ca e ake - epo ed se e i y o illness, plasma α-1 acid glycop o ein, and C- eac i e p o ein >5 mg/L. A ollow-up a e 8 weeks, median hymus a ea had inc eased o 2.5 cm 2 (p< 0.001). Inc ease in hymus a ea du ing ea men was associa ed wi h simul aneous inc ease in mid-uppe -a m ci cum e ence, wi h 0.29 cm 2 highe inc ease in hymus a ea pe cm la ge inc emen in MUAC (p= 0.03). Child en whose F-75 had pa ially been eplaced by ice po idge du ing hei hospi al admission had less inc ease in hymus a ea a e 8 weeks. Conclusion: Malnu i ion and in lamma ion a e associa ed wi h hymus a ophy, and hymus a ea seems posi i ely associa ed wi h plasma phospha e. Subs i u ing he apeu ic o mula wi h un o i ied ice po idge wi h he aim o alle ia ing dia hea may impai egain o hymus size wi h nu i ional ehabili a ion. This calls o esea ch in o possible e ec s o phospha e s a us on hymus size and o he immunological ma ke s. T ial egis a ion: The s udy is based on da a om he FeedSAM s udy, ISRCTN55092738. Keywo ds: Re- eeding, Immune unc ion, Unde nu i ion, Thymus, Elec oly es, In lamma ion, Phospha e * Co espondence: [email p o ec ed] 1 Depa men o Nu i ion, Exe cise and Spo s, Uni e si y o Copenhagen, Roligheds ej 30, 1958 F ede iksbe g C, Denma k Full lis o au ho in o ma ion is a ailable a he end o he a icle © The Au ho (s). 2017 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. Ry e e al. BMC Pedia ics (2017) 17:70 DOI 10.1186/s12887-017-0821-0 Backg ound Se e e acu e malnu i ion (SAM) in child en is a li e- h ea ening condi ion [1], and many hospi als in sub- Saha an A ica epo mo ali y a es abo e 20% in child en admi ed wi h SAM [2]. The easons o he high mo ali y a e no clea , bu mos dea hs a e a ibu able o in ec ious diseases, p obably acili a ed by impai ed immune unc ion in he malnou ished child en [3]. Al hough he mechanism behind he immune de iciency o malnu i ion is s ill poo ly unde s ood, one immunological al e a ion consis en ly e- po ed in malnou ished child en is a ophy o he hymus [4–6]. As such, he size o he hymus has been sugges ed o be a ma ke o he immunode iciency o malnu i ion [3]. E en in absence o se e e acu e malnu i ion, hymus size is associa ed wi h nu i ional s a us, and independen ly o nu i ional s a us, child en wi h a small hymus ha e highe mo ali y [7, 8]. Al hough i is unknown o which ex end hymus size e lec s immune compe ence, he ob- se a ions sugges ha hymus size may be a ma ke o o- bus ness in small child en. This could po en ially make i highly ele an o s udy in child en wi h SAM, who a e ex- emely ulne ablein he i s place. E en hough malnou ished child en a e known o ha e hymus a ophy, i is unknown how his is modi ied by clinical ac o s, such as edema, anemia, elec oly e dis u bances, Human Immunode iciency Vi us (HIV), o he in ec ions o in lamma ion in gene al. Fu he - mo e, al hough p e ious s udies ha e documen ed ha hymus a ophy is e e sible when malnu i ion is ea ed [5, 6], li le is known abou wha de e mines hymus g ow h wi h nu i ional ehabili a ion. The aim o his s udy was he e o e o in es iga e clinical ac o s associa ed wi h hymus size in child en admi ed o in- hospi al ea men o SAM, and p edic o s o g ow h in hymus size wi h nu i ional ehabili a ion. Me hods S udy design This s udy was an obse a ional s udy among child en admi ed o in-hospi al ea men o SAM be ween Oc obe 2012 and Janua y 2013. The s udy was nes ed wi hin he FeedSAM s udy, in es iga ing physiological changes in child en hospi alized wi h SAM, p ima ily change in plasma phospha e (P-phospha e). The Feed- SAM s udy was egis e ed in he ISRCTN egis y wi h he numbe ISRCTN55092738. S udy si e and s anda d ea men Mwanamugimu Nu i ion Uni a Mulago Hospi al is he main ea men cen e o child en wi h complica ed SAM in Uganda. A he ime o he s udy, all child en ecei ed in-pa ien ea men based on he Ugandan Na ional P o ocol o he In eg a ed Managemen o Acu e Malnu i ion, using Wo ld Heal h O ganiza ion (WHO)- ecommended milk-based die s, F-75 and F-100 (Nu ise , F ance), as well as empi ic pa en e al an ibi- o ics, usually ampicillin and gen amycin [9]. Dehyd a ion was ea ed wi h o al ehyd a ion solu ion o mal- nou ished child en (ReSoMal, Nu ise , F ance). When child en we e clinically well hey we e discha ged o ou - pa ien ea men wi h eady- o-use he apeu ic ood [10]. All biological mo he s we e o e ed ou ine coun- seling and es ing o HIV an ibodies, and i he mo he was posi i e o absen , he child was es ed. An ibody- posi i e child en aged <18 mon hs we e e e ed o wi h PCR-based es ing, acco ding o WHO guidelines [11]. Dia hea is a majo conce n a he uni , and a he ime o he s udy, F-75 o F-100 we e occasionally eplaced wi h un o i ied ice po idge o some days, when child en had o de eloped dia hea, and in ole ance o he milk-based eeds was suspec ed. Inclusion and exclusion c i e ia Inclusion c i e ia o child en we e: age 6–59 mon hs; admission on weekdays o ea men o SAM, de ined as ei he weigh - o -leng h z-sco e (WLZ) < −3, using WHO G ow h S anda d [12], o mid-uppe a m ci cum e ence (MUAC) <11.5 cm, o bila e al pi ing edema; li ing close o he hospi al; and a gua dian p o iding in o med con- sen . Exclusion c i e ia we e: signi ican disabili y; mani es shock o se e e espi a o y dis ess equi ing esusci a ion a admission; hemoglobin <4 g/dl o a body weigh <4.5 kg. Se e e in ec ions such as sepsis, HIV o ube culosis we e no easons o exclusion. Inclusion and ollow-up mea- su emen swe eonlypossiblewhenMJHRwasp esen o pe o m he ul asound measu emen s. Da a collec ion A admission, we ob ained in o ma ion abou he child’s cu en symp oms and his o y using a s uc u ed ques- ionnai e. Ca e ake s we e asked o a e he pe cei ed se e i y o hei child’s illness on a isual analogue scale (VAS) om 1 o 10. Vi al signs we e no ed (axilla y empe a u e, pulse, espi a o y a e, and capilla y e ill ime), as well as edema and o al h ush. To assess appe- i e, we no ed whe he he child was able o consume all o he i s se ed he apeu ic eed. Body weigh was measu ed daily on a digi al scale, o he nea es 100 g. Leng h and MUAC we e measu ed o he nea es one mm, using an in an leng h boa d, and measu emen ape, espec i ely. Fo analysis, an h opo- me ic z-sco es we e compu ed using WHO G ow h S anda ds [12]. In o de o ob ain he “ ue”body weigh , a e loss o edema, we used he lowes weigh eco ded a e admission. We ook in o accoun ha leng h was measu ed in all child en by sub ac ing 0,7 cm om leng h measu emen s in child en olde han 2 yea s. Ry e e al. BMC Pedia ics (2017) 17:70 Page 2 o 12 S udy s a moni o ed child en daily (on weekdays), eco ding weigh , equency and consis ence o s ools, ype and amoun o eed gi en, whe he ReSoMal was gi en, and whe he a naso-gas ic ube was used o eeding. Child en we e classi ied as ha ing dia hea when passing h ee o mo e loose o wa e y s ools pe day. Thymus a ea measu emen The same in es iga o (MJHR) measu ed hymus a ea a h ee ime-poin s in all he child en: One o he i s days a e admission (usually day 1 o 2), a ew days be o e discha ge om hospi al, and a ollow-up, app oxima ely 8 weeks a e admission (Fig. 1). MJHR had ained wi h a pedia ic ca diologis expe ienced in hymus ul asound (DLJ) a Copenhagen Uni e si y Hospi al H ido e in Denma k. DLJ supe ised he da a collec ion by e iewing selec ed ul asound images send by email. Thymus size was measu ed using a po able ul asound de ice (Mic oMax, SonoSi e, USA) wi h a pedia ic abdominal p obe. The child was lying on he back, in a bed o on he mo he s lap, and he ansduce was placed on he child’s ches , o e he s e nal bone, in a sagi al p ojec ion h ough he ches (Fig. 2). The hymus was iden i ied as an echo-poo homogenous s uc u e in he medias inum, an e io o, and a ound he g ea essels and he hea (Fig. 3). The la ges lobe o he hymus was iden i ied, and he a ea measu ed. Up o h ee a ea measu emen s we e ob ained pe in es iga- ion, and he a e age calcula ed. In some cases, i was no possible o isualize he hymus. Blood sampling Blood samples ele an o his s udy we e collec ed om a pe iphe al ein a h ee ime poin s: Sample one a admission be o e s a ing e eeding; sample wo app oxi- ma ely 48 h a e s a ing e eeding; and sample h ee, a he day o discha ge o ou pa ien ea men . On all h ee occasions, 1 ml was collec ed in hepa inized e acua ed ubes, and on admission and discha ge, 5 ml was also collec ed in a Cell P epa a ion Vacu aine ® wi h ci a e (Bec on Dickinson, USA). A admission and discha ge, hemoglobin le el was measu ed in hepa inized ull blood, using HemoCue® (Hb 201+, Ängelholm, Sweden). Ci a e plasma was ozen a −80 C o , and shipped o Denma k on d y ice, whe e C- eac i e p o ein (CRP) and α 1 -acid-glycop o ein (AGP), we e measu ed a Uni e si y o Copenhagen, De- pa men o Nu i ion, Exe cise and Spo s, using ABX Pen a® 400 (HORIBA, F ance). Hepa inized plasma om all h ee ime poin s was ozen o −20 C o o up o 2 mon hs, and ino ganic phospha e (P-phospha e) was measu ed a Ebeneze L d Clinical Labo a o y in Kampala (ISO 15189, Labo a o y No. M0221), using molybda e UV me hod (Cobas In eg a® 400 Plus). Con ol g oup Appa en ly heal hy child en, wi h WLZ > −1 and aged 6–59 mon hs, we e ec ui ed among child en o hos- pi al s a and siblings o hospi alized child en and examined once. Thymus a ea assessmen , physical examina ion, an h opome ic measu emen s and blood sampling was done in con ols, as desc ibed o he s udy pa ien s. S a is ics Da a we e en e ed in o EpiDa a (Odense, Denma k) and analyzed using S a a e sion 12 (S a aCo p LP, College s a ion, Texas, USA). No mally dis ibu ed a iables we e exp essed as means ± s anda d de ia ions (SD), and a iables ha did no ollow a no mal dis ibu ion we e exp essed a medians and in e qua ile anges (IQR). Two-sample - es s we e used o e alua e di e ences in means excep in case o non-no mally dis ibu ed ou - comes whe e Mann-Whi ney ank-sum- es s we e used. Chi-squa e es s we e used o compa e p opo ions, excep when he expec ed numbe s we e less han i e, in which case Fishe ’s exac es was used. Fig. 1 O e iew o assessmen s in s udy (Hb: Haemoglobin; CRP: C- eac i e p o ein; AGP: Alpha-1 acid glycop o ein) Ry e e al. BMC Pedia ics (2017) 17:70 Page 3 o 12 To iden i y co ela es o hymus a ea on admission, while including child en wi h an in isibly small hymus, we used an analysis o co a iance (ANCOVA) allowing o le -censo ed measu emen s om child en wi h an in isibly small hymus. Since a ea measu emen s <1 cm 2 may be less accu a e, we assigned all child en wi h an unde ec able hymus and child en wi h measu ed a ea <1 cm 2 o an unknown low alue <1 cm 2 , and such le —censo ed measu emen s ecei ed less weigh in he analysis as compa ed o accu a ely obse ed measu e- men s, using he command “ obi ”in S a a. Thymus a ea did no ollow a no mal dis ibu ion, and hence was loga i hm- ans o med (base 10); es ima es we e subse- quen ly back- ans o med. The analysis was adjus ed o age and sex. ANCOVA was also used o e alua e p edic o s o g ow h in hymus a ea a discha ge and ollow-up, de- ined as he change in hymus a ea (Δ hymus a ea = hy- mus a ea a ollow-up – hymus a ea a admission). These analyses we e adjus ed o age, sex, numbe o days since i s scan and hymus a ea on admission. Child en wi h an unde ec able hymus on admission we e assigned a hymus a ea o 1 cm 2 in o de o calcu- la e he change in hymus a ea o e ime. As sensi i i y analyses, we i s ly analyzed co ela es o hymus a ea and o p edic o s o g ow h in hymus a ea while only including child en wi h a isible hymus, and wi hou le -censo ing o low alues. Secondly, o assess i he associa ions we e explained by body size, we ana- lyzed co ela es o hymus size while adjus ing o body weigh a e loss o edema. Resul s O 120 child en included in he FeedSAM s udy, 85 (71.7%) we e included in he sub-s udy o hymus size measu emen (Fig. 4). O he child en no included in his s udy, 29 we e admi ed on days on which he pe son pe o ming he ul asound scans was no p esen a he uni , 4 child en died be o e ul asound scans we e pe o med, one was excluded be o e scanning due o a hemoglobin le el <4 g/dl, and ano he child was ex- cluded due o a suspec ed medias inal lymphoma. The included 85 child en had a median age o 16 mon hs (IQR: 13; 23 mon hs), 27 (32%) we e gi ls, and 53 (62%) p esen ed wi h edema ous malnu i ion. HIV s a us was unknown in 8 child en (9%), and among he emaining, 15 (19%) we e ound o be HIV in ec ed. O he indings om he ull coho o child en ha e been desc ibed elsewhe e [13–19]. In 22 (26%) o 85 child en scanned on admission, he hymus was no isible by ul asound. Child en wi h an unde ec able hymus had been a ed mo e sick by hei ca e ake s, we e less likely o comple e hei i s he a- peu ic eed (53 s. 80%, p= 0.02), and had highe AGP (2.73 s. 2.24 g/L, p= 0.02) (Table 1). P-phospha e was lowe wo days a e admission in child en wi h an un- de ec able hymus (1.30 s. 1.55 mmol/L, p= 0.01). Respi a- o y symp oms (high espi a o y a e o cough) did no di e in child en wi h and wi hou a de ec able hymus. O he 85 child en scanned a admission, 54 we e scanned a discha ge, meaning ha 31 we e los be o e discha ge: 13 child en died, 13 sel -discha ge be o e ecommended, and i e we e discha ged when he pe - son pe o ming he scans was no p esen a he uni . Be o e ollow-up, ano he 20 child en we e los om he s udy; 5 because hey did no show up a ollow-up, and 15 because he examine was no p esen on he day o ollow-up, lea ing 34 wi h comple e da a a all h ee ime poin s o he s udy (Fig. 4). A g ea e p opo ion o hose los o ollow-up we e boys, while hey did no di e signi ican ly in e ms o an h opome y, age, HIV-in ec ion o acu e-phase eac- an s. Mo e o hose los o ollow-up had an unde ec - able hymus on admission (35% s. 12%, p= 0.02), bu among hose whe e i could be measu ed, hymus a ea was no di e en om hose emaining in he s udy. Fig. 2 Measu ing hymus size using ul asound in a child Fig. 3 Ul asound image o hymus in a malnou ished child in he sagi al iew. The line on he image aces he ou line o he hymus, o measu e he a ea Ry e e al. BMC Pedia ics (2017) 17:70 Page 4 o 12 Co ela es o hymus size on admission Among malnou ished child en wi h a isible hymus, hymus a ea anged om 0.6 o 5.4 cm 2 , wi h a median o 1.3 cm 2 on admission. In con ol child en, hymus a ea anged om 2.4 o 5.3 cm 2 , wi h a median o 3.5 cm 2 (p< 0.001). Thymus a ea co ela ed posi i ely wi h mos an h opo- me ic indica o s, including body weigh , MUAC, weigh - o -age z-sco e, WLZ, and leng h- o -age (Table 2). Ca e ake -pe cei ed se e i y o illness measu ed on he VAS scale co ela ed nega i ely wi h hymus a ea, and so did AGP (10 β = 0.73, 95% con idence in e al (CI): 0.60; 0.90), meaning ha hymus a ea was 27% lowe pe each g/l inc ease in AGP. Child en wi h CRP > 5 mg/l had a 39% smalle hymus a ea han hose wi h lowe CRP (CI: −59%; −11%). The hymus was 27% smalle in HIV in ec ed child en, app oaching signi icance (p= 0.06). Thy- mus a ea co ela ed wi h hemoglobin, and ma ginally wi h P-phospha e on he day o admission. On day wo a e admission, he co ela ion be ween P-phospha e and hymus a ea was s onge , wi h a 66% bigge hymus a ea pe mmol/l highe P-phospha e (CI: 24%; 220%). In he i s sensi i i y analysis, only including child en wi h a isible hymus, we ound o e all simila associa- ions. In he second sensi i i y analysis, adjus ing o body weigh , age and sex, he associa ions we e also simila , excep he e was no longe any signi ican associa ion wi h WLZ o weigh - o -age z-sco e (no shown in ables). Inc ease in hymus a ea Child en we e admi ed o a median o 16 days. Median hymus a ea inc eased om 1.3 o 1.6 cm 2 (p= 0.006) a discha ge, and o 2.5 cm 2 (p< 0.001) a ollow-up (Table 3). These igu es we e i ually unchanged when es ic ing he analysis o only child en who we e ollowed h oughou he s udy. Despi e he g ow h, hymus a ea a ollow-up was s ill signi ican ly smalle han hymus a ea in heal hy child en (p< 0.001). While he hymus was unde ec - able in 22 (26%) child en on admission, his was he case o se en (13%) o 53 child en a discha ge, bu in no child en on ollow-up, o in heal hy con ols. Few ac o s we e associa ed wi h inc ease in hymus a ea du ing nu i ional ehabili a ion (Table 4). Hemoglobin le el measu ed on admission was posi i ely associa ed wi h in- c ease in hymus a ea a discha ge, and AGP was nega i ely associa ed wi h g ow h in hymus a ea a ollow-up. The only an h opome ic indica o associa ed wi h hymus g ow h was inc ease in MUAC, wi h 0.20 cm 2 highe inc ease in hymus a ea pe cm highe inc ease in MUAC a discha ge (CI: 0.05 cm 2 ;0.36cm 2 ), and a ollow-up (0.29 cm 2 highe inc ease in hymus size pe cm highe inc ease in MUAC, CI: 0.04 cm 2 ;0.54cm 2 ). Child en gi en un o i ied ice Fig. 4 Flow diag am showing pa ien s included and assessed a each ime poin Ry e e al. BMC Pedia ics (2017) 17:70 Page 5 o 12 po idge in hospi al had a 0.67 cm 2 smalle inc ease in hymus a ea a ollow-up (CI: −1.28 cm 2 ;−0.05 cm 2 ). The sensi i i y analyses showed simila associa ions, only in- cluding child en wi h a isible hymus on admission. Discussion Thymus a ophy has p e iously been epo ed among malnou ished child en, based on au opsy s udies [4] and in ul asound s udies [5, 6]. I has been hypo hesized Table 1 An h opome ic, clinical and biochemical cha ac e is ics o 85 child en admi ed wi h se e e acu e malnu i ion by isibili y o he hymus a n b Thymus isible Thymus no isible P n=63 n=22 Female sex 85 37 (23) 18 (4) 0.11 Age, mon hs 85 16.0 (12.8; 22.7) 17.3 (13.2; 22.7) 0.59 Edema 85 62 (39) 64 (14) 0.88 S ill b eas eeding 80 17(10) 14(3) 1.00 An h opome ic da a Mid-uppe a m ci cum e ence, cm 84 11.7 ± 1.4 11.6 ± 1.2 0.84 Weigh - o -leng h z-sco e c 85 −3.4 ± 1.5 −3.5 ± 1.3 0.75 Weigh - o -age Z-sco e c 85 −3.9 ± 1.2 −3.9 ± 1.1 0.92 Leng h- o -age Z-sco e 85 −3.1 ± 1.4 −3.0 ± 1.4 0.85 Clinical da a HIV posi i e 77 16 (9) 32(6) 0.18 Symp oms epo ed by ca e ake Dia hea 81 42 (25) 45 (10) 0.80 Vomi 81 39 (23) 55 (12) 0.21 Cough 81 54 (32) 68 (15) 0.75 Fe e 81 27 (22) 37 (6) 0.40 How sick acco ding o ca e ake d 80 6.4 ± 1.7 7.5 ± 2.2 0.03 Physical examina ion Pulse, B/min 81 138 ± 22 138 ± 21 0.95 Respi a o y a e 80 38 ± 11 37 ± 10 0.67 Capilla y e ill ime, sec 82 1.9 ± 0.8 2.1 ± 1.1 0.33 Tempe a u e >37.5 °C 83 24 (15) 25 (5) 0.94 O al h ush 71 26 (13) 29 (6) 0.82 Able o comple e i s eed 73 80 (43) 53 (10) 0.02 Biochemical da a e Hemoglobin, g/dL 80 9.1 ± 2.3 8.7 ± 2.3 0.42 C- eac i e p o ein, mg/L 65 19.4 (7.9; 37.3) 19.6 (12.8; 23.8) 0.19 > 5 mg/L 65 83 (40) 94 (16) 0.43 α 1 -acid glycop o ein, g/L 65 2.24 ± 0.72 2.73 ± 0.66 0.02 Sodium, mmol/L 81 138 ± 4 140 ± 5 0.03 Po assium, mmol/L 81 4.2 ± 0.7 4.0 ± 0.9 0.47 Ino ganic phospha e, mmol/L Admission 82 1.07 ± 0.31 1.03 ± 0.29 0.63 Day wo 72 1.55 ± 0.37 1.30 ± 0.35 0.01 Change du ing i s wo days 70 0.50 ± 0.38 0.31 ± 0.31 0.07 a Values p esen ed a e % (N), median (25%; 75%), o mean ± SD; Di e ences a e conside ed signi ican when p< 0.05 b Numbe o child en wi h ac o s eco ded c Using lowes weigh eco ded du ing admission, a e loss o edema d epo ed on a Visual Analogue Scale om 0 = pe ec ly heal hy, o 10 = as sick as imaginable e all alues excep hemoglobin measu ed in plasma Ry e e al. BMC Pedia ics (2017) 17:70 Page 6 o 12 ha hymus a ophy could e lec he immune de iciency o malnu i ion, causing g ea e suscep ibili y o in ec ions in malnou ished child en [3]. In communi y s udies o child en om Guinea Bissau and Bangladesh [7, 8] chil- d en wi h a small hymus had a highe mo ali y isk, indi- ca ing ha hymus size could be a ma ke o immune compe ence, o pe haps jus a ma ke o good heal h o obus ness. Howe e , ac o s de e mining hymus size in child en wi h SAM, o hymus g ow h wi h nu i ional ehabili a ion, ha e no p e iously been epo ed. Using hymus a ea as a ma ke , we ound ha hymus size was posi i ely associa ed wi h mos an h opome ic Table 2 Linea eg ession iden i ying ac o s associa ed wi h hymus a ea a on admission among 85 child en admi ed wi h se e e acu e malnu i ion. In isible hymuses and alues <1 cm 2 censo ed as “below de ec ion limi ” b n c 10 β (95% con idence in e al) p Female sex 85 1.18 (0.90; 1.54) 0.22 Age, mon hs 85 0.99 (0.98; 1.01) 0.51 Edema p esen 85 1.22 (0.93; 1.58) 0.15 S ill b eas eeding 80 1.09 (0.93; 1.58) 0.64 An h opome ic da a Mid-uppe a m ci cum e ence, cm 84 1.10 (1.01;1.19) 0.02 Weigh - o -leng h, z-sco e d 85 1.11 (1.01; 1.21) 0.03 Weigh - o -age, z-sco e d 85 1.13 (1.02; 1.25) 0.02 Leng h- o -age, z-sco e 85 1.07 (0.97; 1.19) 0.18 Clinical da a, admission HIV posi i e 77 0.73 (0.53; 1.01) 0.06 Symp oms epo ed by ca e ake Dia hea 81 0.94 (0.71;1.22) 0.62 Vomi 81 0.86 (0.65; 1.13) 0.27 Cough 81 0.86 (0.66; 1.12) 0.27 Fe e 81 0.97 (0.73; 1.29) 0.83 How sick acco ding o ca e ake e 80 0.89 (0.83; 0.95) 0.001 Physical examina ion on admission Pulse, bea s/minu e 81 1.00 (0.99; 1.00) 0.10 Respi a o y a e, b ea hs/minu e 80 1.00 (0.99; 1.00) 0.59 Capilla y e ill ime, seconds 82 0.97 (0.84; 1.12) 0.68 Tempe a u e >37.5° 83 0.80 (0.60; 1.08) 0.14 O al h ush p esen 71 0.88 (0.98; 1.69) 0.43 Able o comple e i s eed 73 1.19 (0.86; 1.66) 0.29 Blood chemis y on admission Hemoglobin, g/dL 80 1.08 (1.02; 1.14) 0.01 C- eac i e p o ein > 5 mg/L 65 0.61 (0.41; 0.89) 0.01 α 1 -acid glycop o ein, g/L 65 0.73 (0.60; 0.90) 0.003 Sodium, mmol/L 81 0.99 (0.96; 1.02) 0.49 Po assium, mmol/L 81 1.01 (0.85; 1.21) 0.89 Ino ganic phospha e, mmol/L Admission 81 1.50 (0.99; 2.27) 0.05 Day wo 72 1.66 (1.24; 2.20) 0.001 a Thymus size is log 10 o hymus a ea b Da a a e back- ans o med eg ession coe icien s, adjus ed o age and sex. In e p e a ion o e.g. 10 β = 1.08 is ha by each uni inc ease in exposu e a iable, hymus a ea inc eases by 8%; Associa ions a e conside ed signi ican when p< 0.05 c Numbe o child en wi h ac o s eco ded d Using lowes weigh eco ded du ing admission, o accoun o loss o oedema e E alua ed on a isual analogue scale om 1 o 10 All alues excep hemoglobin measu ed in plasma Ry e e al. BMC Pedia ics (2017) 17:70 Page 7 o 12 indica o s o nu i ional s a us. Simila obse a ions ha e been done among appa en ly heal hy child en in Guinea Bissau [7], Gambia [20] and Bangladesh [21], and hey con i m he concep o he hymus being a “ba ome e o malnu i ion”[22]. We ound ha hymus size was nega i ely associa ed wi h acu e phase eac an s in plasma. This could bo h e lec acu e in ec ions and ch onic, low-g ade in lamma ion. P e ious s udies ha e sugges ed ha acu e in ec ions such as mala ia [7], and neona al in ec ions [23] can cause hymus a ophy, and in lamma ion has ecen ly been epo ed as a majo isk ac o o dea h in child en wi h se e e malnu i ion [24]. The ac ha hymus size was educed in child en wi h CRP abo e jus 5 mg/l could sugges ha low-g ade ch onic in lamma ion could also educe hymus size, al hough dis inguishing be ween in ec ions and in lam- ma ion may be somewha specula i e. Ou s udy con- i ms ha hymus size is educed by nu i ional insul s and in ec ions. Bo h cause ele a ed le els o co isol, which animal s udies ha e ound o cause hymus a ophy [25], simila o low le els o lep in [26]. We saw a nega i e associa ion wi h ca e-gi e epo ed se e i y o illness, simila o a s udy among newbo n child en in Guinea Bissau indica ing ha child en who we e “no well” acco ding o hei mo he s, had a smalle hymus, and sug- ges ed hymus size o be a gene al “ba ome e o good heal h”[27]. Al hough malnu i ion has been associa ed wi h o he immune abno mali ies, like al e ed lymphocy e numbe s and unc ion, in child en [28], as well as in animal expe imen s [29], i is s ill no known whe he hymus size is ac ually linked o immune unc ion, o whe he he asso- cia ion be ween hymus size and isk o dying is caused by o he non-immunological con ounding ac o s, such as diagnosed o undiagnosed in ec ions. As child en admi ed wi h SAM a e o en bo h in- ec ed and malnou ished, i is no su p ising ha 27% o child en in ou s udy had an unde ec able hymus on ad- mission. Simila ly, an au opsy s udy o se e ely malnou - ished child en epo ed how hei hymus was educed o “an i egula s and o ib ous issue”[4]. Ul asound in isibili y o he hymus could also be caused by hype -in la ed lungs, (caused by e.g. pneumonia), p e- en ing ul asound pene a ion. Howe e , nei he e- po ed cough, no espi a o y a e was signi ican ly di e en in child en wi h o wi hou a isible hymus, sugges ing ha pneumonia may no be an impo an Table 3 Thymus size and o he cha ac e is ics in child en du ing ea men o se e e acu e malnu i ion, and in a g oup o heal hy child en a Malnou ished child en du ing ea men Heal hy child en Admission Discha ge Follow-up No. o child en scanned 85 54 34 20 No. o child en wi h isible hymus 74 (63) 87 (47) 100 (34) 100 (20) Thymus a ea, cm 2b 1.3 (1.0; 1.7) 1.6 (1.4; 2.1) 2.5 (2.1; 3.3) 3.5 (3.1; 3.8) Time om admission scanned, days 1 (1; 1) 14 (12; 20) 56 (50; 57) - Time admi ed, days - 16 (13;22) - Weigh gain, g/kg/day - 5.9 (3.7; 8.3) 4.7 (3.0; 6.4) - Weigh , kg 6.8 ± 1.5 7.6 ± 1.4 8.5 ± 1.3 11.2 Leng h, cm 72.7 ± 5.7 72.2 ± 5.2 73.6 ± 5.6 80.7 ± 9.3 Weigh - o -age, z-sco e c - 3.9 ± 1.2 - 3.1 ± 1.1 −2.2 ± 1.0 0.0 ± 1.0 Leng h- o -age, z-sco e - 3.1 ± 1.4 −3.4 ± 1.2 −3.2 ± 1.1 −0.9 ± 1.2 Weigh - o -leng h, z-sco e c - 3.4 ± 1.4 - 1.78 ± 1.1 −0.7 ± 0.9 0.6 ± 0.9 <−2 and > −3 23 (20) 22 (12) 9 (3) 0 (0) <−3 60 (52) 17 (9) 0 (0) 0 (0) Mid-uppe a m ci cum e ence, cm 11.6 ± 1.4 12.0 ± 1.1 12.9 ± 1.1 14.9 Boys 66 (69) 65 (35) 56 (19) 55 (11) Age, mon hs 16.0 (13.0; 22.7) 16.7 (13.5; 23.9) 17.8 (14.9; 26.0) 20.6 (12.0; 34.4) Cu en ly b eas eeding 16 (13) - - 50 (10) Hemoglobin, g/dl 9.0 ± 2.3 9.7 ± 1.9 - 10.2 ± 1.5 Plasma C- eac i e p o ein, mg/L 19.6 (8.8; 31.2) 0.4 (0.2; 1.9) - 0.8 (0.2;2.8) Plasma α 1 -acid glycop o ein, g/L 2.37 ± 0.73 1.09 ± 0.39 - 0.83 ± 0.30 a Values p esen ed a e n, %(n), median (25%; 75%) o mean ± SD b Only including child en wi h a isible hymus c Admission z-sco es we e compu ed o all child en based on he lowes weigh eco ded (a e loss o edema) Ry e e al. BMC Pedia ics (2017) 17:70 Page 8 o 12 cause in his coho . I is possible ha he hymus in some cases would ha e been isualized in he hands o a mo e expe ienced sonog aphe . Howe e , he ac ha we saw simila associa ions when including in isible hymuses as “in isibly small”, and when including only isible hymuses (in he sensi i i y analysis) sugges s ha i may be easonable o assume ha hymuses we e unde ec able because hey we e e y small. Thymus size was posi i ely associa ed wi h hemoglobin le el, and P-phospha e, which o ou knowledge, has no p e iously been epo ed. The g ea e associa ion wi h phospha e on day wo may be explained by he ac ha mos ul asound scans we e done one o wo days a e admission, and he e o e close in ime o he second blood sample. In ec ions may cause bo h anemia and hypophospha emia [30], and hus he associa ion could e lec he e ec o in lamma ion on hymus size. How- e e , he associa ions pe sis ed a e adjus ing o CRP and AGP. Anemia and hypophospha emia may be ma ke s o poo nu i ional s a us, al hough hei associa ions wi h hymus size emained in he sensi i i y analysis adjus ing o body size. I is also plausible ha hypophospha emia by i sel may con ibu e o hymus a ophy. Hypophospha- emia has been ound o cause leukocy e dys unc ion [31], and phospho us is a ype II nu ien , essen ial o g ow h and main enance o lean body mass [32]. Thymus a ophy occu s in animals de icien in o he ype II nu ien s, like zinc [33] and magnesium [34], and a ecen s udy ound Table 4 Co ela es o change in hymus size om admission o discha ge and o ollow-up among child en ea ed o se e e acu e malnu i ion a To discha ge To ollow-up n b β(95%CI) p n b β(95% CI) p Female sex 47 0.00 (−0.30;0.31) 0.98 34 −0.06 (−0.65; 0.53) 0.84 Age, mon hs 47 0.01 (−0.01; 0.03) 0.40 34 - 0.02 (−0.06; 0.01) 0.19 Days om admission 47 0.02 (−0.01; 0.04) 0.21 34 0.00 (−0.04; 0.05) 0.94 Clinical da a, admission Edema p esen 47 −0.14 (−0.50;0.23) 0.46 34 −0.34 (−0.97; 0.29) 0.28 HIV in ec ed 46 0.09 (−0.38; 0.55) 0.71 33 0.35 (−0.60; 1.31) 0.45 S ill b eas eeding 43 0.08 (−0.38;0.54) 0.71 31 0.25 (−0.66; 1.17) 0.58 How sick acco ding o ca e ake c 46 0.04 (−0.05; 0.13) 0.35 34 0.07 (−0.13;0.28) 0.46 Physical examina ion, admission Tempe a u e >37.5° 46 0.01 (−0.43; 0.45) 0.96 34 0.10 (−0.66; 0.86) 0.79 Capilla y e ill ime, sec 45 −0.05 (−0.24; 0.14) 0.62 32 0.14 (−0.34; 0.62) 0.55 Able o comple e i s eed 44 −0.04 (−0.40; 0.32) 0.81 31 0.35 (−0.45; 1.15)) 0.37 Blood chemis y, admission C- eac i e p o ein >5 mg/L 38 0.36 (−0.07; 0.79) 0.10 26 −0.22 (−1.16; 0.72) 0.63 α 1 -acid glycop o ein, g/L 38 0.03 (−0.24; 0.31) 0.81 26 −0.60 (−1.12; −0.08) 0.03 Hemoglobin, g/dL 46 0.08 (0.01; 0.15) 0.02 33 −0.00 (−0.14; 0.13) 0.95 Ino ganic phospha e, mmol/L 46 0.16 (−0.34; 0.66) 0.52 33 0.12 (−1.01; 1.24) 0.83 An h opome ic g ow h in same pe iod Weigh gain a e, kg/day d 47 5.77 (−1.20; 12.74) 0.10 34 14.02 (−8.12; 36.15) 0.21 Δmid-uppe a m ci cum e ence, cm 46 0.20 (0.05;0.36) 0.01 23 0.29 (0.04; 0.54) 0.03 Δweigh - o -leng h z-sco e d 47 −0.02 (−0.20; 0.17) 0.87 34 0.08 (−0.19; 0.34) 0.57 Δweigh - o -age z-sco e d 47 0.08 (−0.25; 0.41) 0.62 34 0.22 (−0.20; 0.63) 0.30 Obse a ions and ea men s gi en du ing admission Dia hea obse ed 47 - 0.10 (−0.44; 0.25) 0.58 34 0.10 (−0.52; 0.71) 0.75 Rice po idge gi en 46 −0.11 (−0.42; 0.19) 0.47 33 −0.67 (−1.28; −0.05) 0.03 Naso-gas ic ube used 47 −0.08 (−0.44; 0.28) 0.67 34 −0.25 (−0.92; 0.42) 0.46 a Da a shown a e eg ession coe icien s o linea eg ession analysis o change in hymus size (Δ hymus size) adjus ed o hymus size on admission, days since admission, age and sex. Child en wi h in isible hymus on admission we e assumed o ha e hymus a ea = 1 cm 2 ; In e p e a ion o e.g. β= 0.20 means a 0.20 cm 2 u he inc ease in hymus size pe uni inc ease in exposu e a iable; Associa ions a e conside ed signi ican when p< 0.05 b n = numbe o child en in whom da a is a ailable c E alua ed on a isual analogue scale om 1 o 10 d Weigh gain = p esen weigh –lowes weigh du ing admission, o accoun o loos o oedema Ry e e al. BMC Pedia ics (2017) 17:70 Page 9 o 12