α‐1 acid glycoprotein but not C‐reactive protein is associated with significantly lower serum zinc concentrations among Congolese children aged 6–59 months and has a substantial impact on prevalence estimates of zinc deficiency

Abstract

BackgroundSerum zinc (Zn) is a negative acute phase reactant; hence, concentrations decrease in the presence of inflammation and infection. There is no current consensus on how to adjust serum Zn concentrations for inflammation.ObjectivesThe aim was to determine associations between serum Zn concentration and inflammation biomarkers (C‐reactive protein [CRP] and α‐1 acid glycoprotein [AGP]) and to compare means and the prevalence of Zn deficiency using unadjusted serum Zn concentrations and serum Zn concentrations adjusted for inflammation using study generated correction factors (CFs) among children in the Democratic Republic of the Congo.MethodsNon‐fasting blood was collected in trace‐element free vacutainers from 744 children (6–59 mo) recruited from South Kivu and Bas Congo provinces in 2014 using a probability proportionate to size sampling method. Serum was analyzed for Zn (n=691), CRP and AGP concentrations (n=687). Linear regression was used to estimate associations between serum Zn and AGP and CRP concentrations and to generate CFs (calculated as 1 divided by the geometric mean ratio) for Zn based on three stages of inflammation: (incubation [CRP >5 mg/L], early convalescence [CRP >5 mg/L and AGP >1 g/L] and late convalescence [AGP >1 g/L]), relative to children with no inflammation.ResultsOverall, the prevalence of acute (CRP >5 mg/L) and chronic (AGP >1 g/L) inflammation was 29% (n=197) and 66% (n=455), respectively. Unadjusted mean (95% CI) serum Zn concentration was 9.4 (9.3, 9.6) μmol/L. Unadjusted mean ± SD serum Zn concentration was 10.0 ± 1.9 μmol/L among children with no inflammation (n=213; 35%), 9.8 ± 1.1 μmol/L among children in the incubation stage (n=11; 2%), 8.7 ± 2.1 μmol/L among children in early convalescence (n=117; 19%), and 9.4 ± 2.1 μmol/L among children in late convalescence (n=260; 43%). A 1 g/L increase in AGP was associated with a 0.5 (95% CI: 0.3, 0.7) μmol/L lower serum Zn concentration (P<0.001). CRP was not significantly associated with serum Zn concentration (P=0.11). Study generated CFs (95% CI) for Zn were 1.01 (0.88, 1.15), 1.16 (1.11, 1.21) and 1.07 (1.03, 1.11) for incubation, early and late convalescence stages, respectively. After applying the CFs, adjusted mean (95% CI) serum Zn concentration was 10.1 (9.8, 10.2) μmol/L. Overall, the prevalence of Zn deficiency (<8.7 μmol/L) decreased from 35% (n=244) using unadjusted serum Zn concentrations to 24% (n=160) using serum Zn concentrations adjusted for inflammation.ConclusionAGP is associated with significantly lower serum Zn concentrations and the application of study generated CFs had a substantial impact on Zn deficiency prevalence rates. AGP appears more useful than CRP to measure the effects of inflammation on serum Zn in our study population; however, we acknowledge that the use of more than one inflammation biomarker is ideal. Adjustment for inflammation appears to be warranted for accurate estimates of population‐level Zn status.Support or Funding InformationFunding for this research was provided by HarvestPlus. C.D.K. received a doctoral research award from the International Development Research Centre (Canada) and a Vanier scholarship from the Canadian Institutes of Health Research (CIHR).