Prematurity is the leading cause of neonatal deaths in Africa. More than one million children die each year due to premature co-morbidities. The conditions that cause the death of these newborns can also result in severe lifelong disabilities for those who survive. Prematurity is a major public health problem in Africa. It is now an indicator of the health of newborns and a key determinant of infant survival, growth and development. In spite of the many proposed medical care protocols, the incidence of prematurity has varied minimally over the past 40 years. In 1998, Tietche et al. estimated the rate of prematurity at 21.05% in the low birth weight unit of the maternity ward at the Yaounde Central hospital [1] . In 2005, Monebenimp et al. reported a prematurity rate of 57% at the Yaoundé University Teaching Hospital with a neonatal mortality rate of 35.8% and a perinatal mortality rate of 92% [2] . Publications on comorbidities associated with prematurity in our context are limited. However, the study by Njom et al. in Yaoundé in 2014 identified the following as the main comorbidities in very premature babies: apnoea, respiratory distress, neonatal infection, anaemia and neonatal jaundice [3] . We set out to determine the incidence of prematurity and co-morbidities in two referral hospitals in Cameroon.

Our study was carried out in the neonatology units of the Douala General Hospital (DGH) and the Laquintinie Hospital (LHD) in Cameroon. The neonatal unit of the Laquintinie Hospital of Douala is classified as a second level hospital in the Cameroon health care structure. However, the neonatal unit of the Douala General Hospital is described as a first grade hospital in the Cameroon healthcare structure, receiving the majority of referrals of ill newborns. These two health units each have a maternity unit and receive newborns between the age of 0 to 1 month from their maternity ward and those of other health units in the city of Douala or Cameroon. It was a hospital-based cross-sectional study, with retrospective data collection conducted from 1 January 2015 to 1 January 2018.

All newborns of less than 37 weeks gestational age admitted to the neonatal service of the DGH and HLD were included. Excluded from our study were: Full-term newborns, any inoperable medical record and newborn discharged against medical advice (DAMA).

We collected data from the patient’s medical records and the hospital registry, based on a pre-established chart. The parameters collected are sociodemographic data (postnatal age, gestational age at birth, address, socio-economic level of the family), maternal and obstetrical data (maternal age, parity, gestation, obstetrical complications, notion of prenatal corticosteroid), childbirth data (place, type of delivery, duration of amniotic sac rupture, appearance of amniotic fluid, adaptation to extra uterine life and neonatal resuscitation ), neonatal pathologies (hyaline membrane disease, apneas, neonatal infections, ulcerative enterocolitis, ....). Data concerning medical care (oxygen therapy, ventilation, antibiotic therapy, phototherapy...) and clinical progress reports (acute complications and lethality). Oxygen therapy was used via oxygen goggles with humidifier, high-concentration oxygen masks or conventional mechanical ventilation, depending on the indication.

Data entry and analysis were performed using Microsoft Excel version 17.2 and the Social Science Statistics package version 20.0.

Inter-variable associations were determined by the Chi² test and the significance level was determined for a value of p < 0.05. A bivariate and then multivariate analysis was carried out using the logistic regression method to identify the factors associated with mortality.

The Kaplan-Meier model was used to estimate the survival rate of the study population.

Bronchopulmonary dysplasia (BPD): This is chronic respiratory failure with persistent oxygen dependence after the 28th day of life in newborns who have presented initial respiratory distress requiring assisted ventilation for at least the first 3 days of life.

Among the 5647 newborns who were admitted to the neonatal units during the study period, we had an incidence of prematurity of 32.5%. Akolly et al. in Togo found similar results in a sample of 566 cases [4] . This value is higher than

those found by Tietche et al. at the central hospital in Yaoundé (21.05%) in 1998 [5] , 8.5% in Nigeria [6] , 11.1% and 4.5% in China [7] . This trend is two or three times higher than that observed in developed countries. In Canada, it was 7.5% [8] . In France from 2016 to 2021, the prematurity rate is stable at 7.0% [9] - [11] . The U.S. preterm birth rate declined 1% in 2022 to 10.38%. This 2022 decline follows a 4% rise in the rate from 2020 to 2021 (10.49%) [12] . This can be explained by the fact that the LHD and the DGH are referral centres in the city of Douala and receive several preterm newborns and high-risk pregnancies from other health facilities that lack adequate equipment and qualified personnel for proper medical care. The majority (53.7 %) of preterm newborns are born between 33 and 36 week’s gestation, with an average gestational age of 32.4 weeks. Likewise, many authors found similar results [13] - [15] . However, these results are different from those observed by some African authors [16] [17] where very preterm newborns were in the majority because their study population was centred on very preterm newborns. The majority of the premature newborns in our study were low birth weight babies (1500 - 2499 g) and represented 65.3% of the population with an average weight of 1747 g.

Intrauterine growth retardation was associated with prematurity in 12.4% of cases. The IUGR rate associated with prematurity varies in the literature from 16.4% to 96.9% [13] [18] - [20] . The association of prematurity and IUGR is common because the etiologies of these two entities are often complex. In our study, the majority of preterm newborns were female (51%), Tietche et al. [5] found a slight female predominance, as did Ndiaye et al. [17] . There are no explanations for this slight predominance of the female sex.

Neonatal infection was found in 86.9% of premature newborns, which is close to the findings of Njom Nlend et al. [3] who reported the incidence of neonatal infections in 80% of newborns in Cameroon. These results are comparable to the prevalence of neonatal bacterial infections in hospitals, documented by positive blood cultures in similar settings [21] . These figures are very high compared to other African studies. Ndiaye et al. had a neonatal infection rate of 64.4% [17] . These differences can be explained by the different criteria (which are not the same) used to diagnose infections. In our context, bacteriological investigations to confirm a possible neonatal infection are occasionally performed because of their high cost, the low socio-economic level of the population and the lack of adequate medical equipment. Antibiotic therapy is sometimes used systematically and wrongly, given the limited availability of diagnostics and the effectiveness of aseptic measures.

Hypothermia was present in 13% of the preterm newborns in our study population on admission. This was less than in the study by Minto’o et al. with 56.5% and represents the clinical aspect which is most frequently associated with prematurity on admission to their unit [21] . The study carried out by Ndiaye et al. in Senegal found that 62.7% of preterm newborns were hypothermic [17] . In our study, we can relate this low rate of hypothermia to the finding that more than half of the preterm deliveries took place in our maternity units located very close to the neonatal units and thus the referral conditions were adequate.

Liver disorders (jaundice and Vitamin K deficiency) are the second most common co-morbidity associated with prematurity, accounting for 54%. Jaundice rates found in our study are higher than the jaundice rate found by Ugochukwu [22] . The elevated rate of liver anomalies in our study could be explained by the fact that our study exclusively involved preterm newborns who are more likely to develop neonatal jaundice secondary to hepatic immaturity. Hematologic disorders ranked third among co-morbidities with 32.9% represented by early secondary anemia at 25.9% and late onset secondary anemia at 8.4%. These data are similar to the study by Njom Nlend et al. with 22% [3] . This can be explained by a low haemoglobin level but also by losses during blood sampling in the first weeks of life. In our study, we observed that only 30.5% of preterm newborns with anaemia have had a blood transfusion and 3% were given erythropoietin. This is attributable to the fact that erythropoietin was not very accessible due to its high cost and that some parents refuse blood transfusions for religious reasons. Respiratory disorders were 13.2%, represented by apnea 10.9% and hyaline membrane disease (HMD) 2.6%. On the other hand, Ikram et al. found 38.6% HMD in preterm infants [13] , Diouf et al. found 16.42% respiratory distress at birth in preterm infants, 43.75% of which was HMD [17] . We can assign this low rate of HMD in our study on the one hand to a diagnostic difficulty since chest X-rays are not performed systematically. On the other hand, there may be an early synthesis of surfactant in the black population (1).

Our study found a mortality rate of 27.4%. Studies carried out in Cameroon revealed a high mortality rate of 37.7% in 2011 and 36.6% in 2013 at the gynaecological-obstetric and paediatric hospital in Yaoundé [23] [24] . In Senegal found 50.3% [17] , Lawn et al. 27% [25] . The low rate in our study may be due to the fact that its calculation is not based on the overall preterm delivery rate (stillbirths and premature live deliveries not included in our study) but only on the live births included in our study.

Among the risk factors, the threatened preterm delivery was one of the independent predictor of mortality. This may be explained by the fact that expectant women after PROM had not received effective antibiotic treatment nor had they received fetal lung maturation therapy (corticosteroids).

Our study found that gestational age ≤ 32 weeks was a predictor of a preterm death. The mortality rate increased as GA decreased. Ugochukwu et al. in Nigeria found a relatively high rate of survivors at 31 to 35 weeks and the majority of deaths at 24 to 29 weeks [22] . Similarly, Ntonya et al. in Malawi found a high mortality rate between 24 and 32 weeks and the majority of survivors between 33 and 37 weeks [26] . This situation could be explained by the lack of adequate medical equipment, in particular the accessibility to ventilators and especially by the unavailability of surfactant in our context.

Place of birth was also significantly associated with preterm mortality. The inadequate medical care of the preterm neonate at birth in these hospital services and the ill-adapted conditions of the preterm neonate during their transfer could have contributed to their death during hospitalisation.

The vast majority of preterm newborns who died had a very low birth weight (VLBW). The relationship between weight and prognosis was not proportional. This is similar to studies by several author [8] [23] [24] . Velaphi et al. in South Africa explained that low birth weight preterm infants had low survival rates because they were not on mechanically assisted ventilation and that artificial ventilation is needed in facilities with limited resources to optimize survival [27] . Low-weight babies are at greater risk of death because of their physical, immunological and biological immaturity to maintain and protect against disease.

The Apgar score < 6, was associated significantly and independently with the death of preterm newborns in our study. Tibaijuka et al. observed that low Apgar score was a factor in mortality [28] . Given that the Apgar score evaluates the vital parameters of the newborn, the vital functions of preterm newborns are often underestimated and, consequently, their survival is also compromised. Although the Apgar score influenced the outcome of the preterm newborn in our study, it is important to recognize its limitations. Late secondary anemia was the only independent predictor associated with death among co-morbidities. This is because the requirements for optimal medical management of premature newborns in our context and in Africa are not yet sufficient compared to the developing world, hence the need for further rapid improvements.

The need for resuscitation during hospitalization was significantly associated with the demise of preterm newborns. The underlying pathology that triggered resuscitation was severe and contributed to the death of the babies, though a majority survived among those resuscitated.

At the end of our study, we can report that the incidence of prematurity during our study period was 32.5% and 10.6% in the Neonatology Units and maternity units respectively. The major comorbidity associated with prematurity was neonatal infection. However, the mortality rate of preterm infants was 27.4%.

Some missing or incomplete data limit the power of the study. Heterogeneity in practices could result from the 2 study sites and could reduce the comparability of the data.

NDD, ED, and DFY designed the study, drafted the initial manuscript, and reviewed and revised the manuscript. MSCO and NDD designed the data collection instruments and collected the data. ED, NS, and MMLE critically reviewed the manuscript for important intellectual content. NDD, MSCO, and MKB coordinated, and supervised data collection. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

We are grateful to all the staff of the neonatal units of Douala General Hospital and Douala Laquintinie Hospital.