Development of locally relevant clinical guidelines for procedure-related neonatal analgesic practice in Kenya: a systematic review and meta-analysis

Background Increasing numbers of neonates are undergoing painful procedures in low-income and middle-income countries, with adequate analgesia seldom used. In collaboration with a multi-disciplinary team in Kenya, we aimed to establish the first evidence-based guidelines for the management of routine procedure-related neonatal pain that consider low-resource hospital settings. METHODS: We did a systematic review by searching MEDLINE, Embase, CINAHL, and CENTRAL databases for studies published from Jan 1, 1953, to March 31, 2019. We included data from randomised controlled trials using heart rate, oxygen saturation (SpO(2)), premature infant pain profile (PIPP) score, neonatal infant pain scale (NIPS) score, neonatal facial coding system score, and douleur aiguë du nouveau-né scale score as pain outcome measures. We excluded studies in which neonates were undergoing circumcision or were intubated, studies from which data were unextractable, or when pain was scored by non-trained individuals. We did a narrative synthesis of all studies, and meta-analysis when data were available from multiple studies comparing the same analgesics and controls and using the same outcome measures. 17 Kenyan health-care professionals formed our clinical guideline development panel, and we used the Grading of Recommendations, Assessment, Development and Evaluation framework and the panel's knowledge of the local health-care context to guide the guideline development process. This study is registered with PROSPERO, CRD42019126620. FINDINGS: Of 2782 studies assessed for eligibility, data from 149 (5%) were analysed, with 80 (3%) of these further contributing to our meta-analysis. We found a high level of certainty for the superiority of breastfeeding over placebo or no intervention (standardised mean differences [SMDs] were −1·40 [95% CI −1·96 to −0·84] in PIPP score and −2·20 [–2·91 to −1·48] in NIPS score), and the superiority of oral sugar solutions over placebo or no intervention (SMDs were −0·38 [–0·61 to −0·16] in heart rate and 0·23 [0·04 to 0·42] in SpO(2)). We found a moderate level of certainty for the superiority for expressed breastmilk over placebo or no intervention (SMDs were −0·46 [95% CI −0·87 to −0·05] in heart rate and 0·48 [0·20 to 0·75] in SpO(2)). Therefore, the panel recommended that breastfeeding should be given as first-line analgesic treatment, initiated at least 2 min pre-procedure. Given contextual factors, for neonates who are unable to breastfeed, 1–2 mL of expressed breastmilk should be given as first-line analgesic, or 1–2 mL of oral sugar (≥10% concentration) as second-line analgesic. The panel also recommended parental presence during procedures with adjunctive provision of skin-to-skin care, or non-nutritive sucking when possible. INTERPRETATION: We have generated Kenya's first neonatal analgesic guidelines for routine procedures, which have been adopted by the Kenyan Ministry of Health, and have shown a framework for clinical guideline development that is applicable to other low-income and middle-income health-care settings. FUNDING: Wellcome Trust Research Programme, and the Africa-Oxford Initiative.