Pengaruh Proses Alkalisasi terhadap Kadar Fenolik Total pada Bubuk Kakao: Kajian Meta-Analisis

Abstract

Pada tahapan pengolahan kakao, proses alkalisasi memengaruhi kadar fenolik total. Penelitian meta-analisis ini bertujuan menganalisis pengaruh tingkat alkalisasi terhadap kadar fenolik total bubuk kakao. Berdasarkan kriteria inklusi, terpilih 8 artikel dengan 32 studi. Effect size dikalkulasi berdasarkan analisis Standardized Mean Difference (SMD) metode Hegdes’d dengan interval kepercayaan 95%. Hasil penelitian menunjukkan proses alkalisasi memberikan efek penurunan kadar fenolik total kakao secara konsisten (SMD: -7.42; 95% CI: -9.94; -4.92) dan signifikan (p-value < 0.05) dengan heterogenitas tinggi antarstudi (I2=71.8%) menandakan intervensi variabel moderator. Analisis subgrup tingkat alkalisasi menunjukkan perlakuan setiap tingkat alkalisasi memberikan efek penurunan konsisten dan signifikan, yakni light (SMD: -5.00; 95% CI: -9.16; -0.85; p-value < 0.0001), medium (SMD: -13.23; 95% CI: -30.58; -4.12; p-value = 0.0008), dan dark (SMD: -16.18; 95% CI: -34.63; -2.27; p-value = 0.0136). Hasil uji one-way ANOVA mengindikasikan peningkatan tingkat alkalisasi dari medium ke dark tidak menghasilkan perbedaan efek yang signifikan secara statistik. Negara manufaktur bubuk kakao dipertimbangkan sebagai variabel moderator, menunjukkan Indonesia (SMD: -109.72; 95% CI: -254.55; -35.11; p-value = 0.0041) dan Kamerun (SMD: -6.57; 95% CI: -11.07; -2.07; p-value = 0.0326) secara berurutan memberikan efek penurunan terbesar dan terendah secara signifikan. Hasil tersebut konsisten dengan uji one-way ANOVA menunjukkan perbedaan nyata antara kelompok negara dengan tingkat penurunan terendah dan tertinggi, khususnya antara Kamerun (subset A) dengan Indonesia (subset C). Penelitian meta-analisis ini didukung hasil robust dan stabil dari bias dengan uji Rosenthal’s Fail-safe Number, yakni nilai fail-safe N melebihi nilai batas minimum 5N+10 dan hasil Egger’s test nilai p > 0.05 mengonfirmasi tidak ada bukti signifikan funnel plot asimetris sebagai tanda minim bias publikasi.

During cocoa processing, alkalization affects the phenolic content of cocoa. This meta-analysis aimed to evaluate the effects of alkalization on total phenolic content. Based on the inclusion criteria, 8 articles comprising 32 studies were selected. Effect sizes were calculated using the Standardized Mean Difference (SMD) method with Hedges’d with a 95% confidence interval. The results demonstrated that the alkalization process consistently and significantly reduced cocoa polyphenol content (SMD: -7.42; 95% CI: -9.94 to -4.91; p-value < 0.05), with high heterogeneity among studies (I² = 71.8%), indicating the influence of moderator variables. Subgroup analysis based on alkalization level showed that each alkalization treatment produced a consistent and significant reduction effect, including light (SMD: -5.00; 95% CI: -9.16 to -0.85; p-value < 0.0001), medium (SMD: -13.23; 95% CI: - 30.58 to -4.12; p-value = 0.0008), and dark (SMD: -16.18; 95% CI: -34.63 to -2.27; p-value = 0.0136). One-way ANOVA indicated the increase in alkalization level from medium to dark did not produce a statistically significant difference in effect. The cocoa-manufacturing country was also considered as a moderator variable, with Indonesia (SMD: -109.72; 95% CI:-254.55 to -35.11; p-value = 0.0041) and Cameroon (SMD: -6.57; 95% CI: - 11.07 to -2.07; p-value = 0.0326) showing consistent and significant reduction effects. One-way ANOVA results revealed significant differences among country groups with the lowest and highest reduction levels, particularly between Cameroon (subset A) and Indonesia (subset C). The findings were further supported by robust and stable results from Rosenthal’s Fail-safe Number analysis, where the fail-safe N exceeded the minimum threshold value of 5N+10. In addition, Egger’s test produced a p-value > 0.05, confirming no significant evidence of asymmetric funnel plot visualization which indicating minimal publication bias.