This study evaluated the trends, causes and associated factors of maternal mortality at a tertiary health facility over a 13-year period (2007–2019), using Joinpoint regression modelling and other statistical methods.
Summary of main findings
We found that over the 13-year study period, 483 maternal deaths occurred at an average of 37 deaths per annum. Furthermore, the MMR increased by 4% per annum from 2211 per 100,000 live births in 2007 to 3556 per 100,000 in 2019. Hypertension, haemorrhage, and sepsis contributed about two-thirds of maternal deaths. Anaemia, HIV/AIDS and Sickle cell disease were the leading indirect causes of maternal death.
Trends in maternal mortality ratio
In contrast to the WHO estimates that suggested that maternal mortality decreased in Nigeria between 1990 and 2015 [
2,
9], our study revealed an overall upward trend in maternal mortality of about 4% per annum from 2007 to 2019. Generally, the MMR of between 1700 and 7000 per 100,000 live births in our institution between 2007 and 2019 was very high as compared to the WHO estimate of 800/100,000 live birth for Nigeria [
2]. Our finding is similar to that of Idoko et al. from Gambia who found that the decreased MMR trends as reported by WHO for Gambia was at variance with their institutional increasing MMR trends [
10]. Likewise, Okonofua et al. found an increased MMR among three secondary health institutions in Lagos state, Nigeria from 2015 to 2016, with similarly high MMR of 1602/100,000 live births [
7]. Our centre is also located in same metropolitan city of Lagos, Nigeria. The poor vital registration system in Nigeria continues to be a challenge and may have negatively impacted on the validity of the WHO estimates of maternal mortality in Nigeria and many other LMICs [
6].
Furthermore, our institutional MMR was comparable to reports from the northern and Eastern parts of Nigeria [
11,
14‐
17]. However, the MMR at our centre was higher than the reports from Calabar, South-south Nigeria [
18] and elsewhere in Nigeria [
4]. International comparisons showed that similarly high MMR was also reported from Gambia, Ghana and some LMICs [
13,
18]. However, our institutional MMR was about 5 to 100-fold higher than values reported from countries in some part of Africa and among Middle-Income and High-Income Countries [
19‐
23]. Our results are in sharp contrast to reports from HICs that had declining or stable trends of MMR [
21‐
25].
However, some other authors also reported slight decline from some part of Nigeria in line with the WHO report [
4,
8,
16‐
18]. The reported decline may be spurious as these reports generally involved shorter study periods of less than ten years which would make it difficult to draw valid conclusion on trends. Furthermore, these trends studies used descriptive rather than analytical statistics.
The high MMR as highlighted by our study may also be partly due to other reasons. Our centre is a tertiary/ referral institution and would attract a higher proportion of high-risk patients and our findings may not be generally applicable. Being a referral centre however should not be a justification for the high and increasing MMR trends as tertiary hospitals are expected to be appropriately equipped and well-staffed to manage complicated maternal health conditions. Poor investment into health infrastructure and inadequate number of skilled and experienced personnel to handle maternal and child health issues at the primary and secondary healthcare facility could have contributed to the increased MMR at our centre, as the inadequate management of cases at the lower level centres may overburden the tertiary hospitals [
3]. Worsening economic outlook, and massive brain drain of health workers continue to accelerate while the effect of the COVID 19 pandemic is yet to be revealed. In order to achieve SDG 3.1 of less than 70 deaths per 100,000 live births by 2030 [
26], global MMR must decrease at a rate of 2% Per annum from 2000 to 2030 SDG. For countries in sub-Saharan Africa, the MMR decline must occur at a much higher rate because the starting figures are well above average. Thus, policies geared towards strengthening institutional response to maternal morbidity and mortality are very important in achieving SDG 3.1 [
7].
Trends in causes of maternal mortality
Nearly two-thirds of maternal mortality at our centre was due to hypertensive disorders, sepsis and haemorrhage. This is the usual pattern in most LMICs [
11,
14‐
17]. Thus, efforts geared towards reducing these major causes of maternal mortality will dramatically reduce maternal deaths at our centre. It is noteworthy that the leading causes of maternal mortality in the United Kingdom for instance is sharply different with cardiac disease, pulmonary embolism, epilepsy being the leading contributors [
27]. As occurred in LMICs, hypertensive disorders were the leading cause of maternal death at our centre (except for the last three years of the study, 2017–2019) [
2,
10,
19]. However, we found that the contribution of hypertensive disorders to overall maternal death in LUTH decreased from 29.9% in 2007–2012 period to 23.6% in 2013–2019 period.
The modest decline in the contributions of hypertensive disorders may be attributable to the use of evidence-based preventive protocols such as the commencement of low dose aspirin and calcium lactate in early pregnancy among high-risk individuals for prevention of preeclampsia [
28] and the introduction and implementation of Magnesium Sulphate in the prevention and management of eclampsia [
28]. Provision of high-quality antenatal care to all pregnant women for early detection of features of preeclampsia, training of peripheral staff to recognise and manage cases of preeclampsia, seamless referral system to tertiary hospital and strict adherence to protocols for the management of hypertensive disorders can further reduce its contributions to maternal deaths in our environment.
Sepsis has been the second leading cause of maternal mortality at our centre for nearly two decades, contributing about one-quarter of maternal deaths in the last seven years (2013–2019) [
3]. However, most studies in LMICs reported sepsis as the third leading cause of maternal mortality after hypertension and haemorrhage [
2,
7,
10,
14,
16,
19]. Prevention of prolonged labour, implementation of asepsis, antiseptic and rational antibiotic protocols and policies across various levels of service delivery can considerably reduce maternal deaths due to sepsis [
3]. Fake and substandard drugs that engender antibiotic resistance might also have contributed to the risk of septicaemia among maternal deaths [
3]. Legislation and enforcement by the relevant authorities will help in mitigating this problem.
During the study period of 2007–2019, haemorrhage was the leading cause of death among the booked patients accounting for one-third of maternal deaths among them. Although haemorrhage was the second leading cause of maternal mortality in most LMICs including Nigeria [
10,
11,
14,
15,
19], haemorrhage became the leading cause of death only in the last two years of our study. Furthermore, the contribution of haemorrhage to annual maternal mortality increased from 11% in 2007–2012 to 19.3% in 2013–2019 study period. Thus, the hospital’s strategy in the management of ante-partum and post-partum haemorrhage must be reviewed. Blood banking system must be further improved and high index of suspicion for identifying risks for post-partum haemorrhage must be improved at the centre. Obstetric drills should be conducted regularly to prepare for eventualities. It is not clear why there was no death due to haemorrhage in 2016 even though haemorrhage contributed 15.3 and 15.0% to maternal deaths in 2015 and 2017 respectively.
Obstetric haemorrhage can be acute and unpredictable and may occur even without any identifiable risk factor [
29]. As previously noted, haemorrhage may not be the leading cause of maternal deaths among un-booked patients possibly because they might have died before getting to our centre [
3].
Indirect causes generally constitute about 25% of maternal death in LMICs [
11,
14,
15,
19]. However, indirect causes contributed about 17% of maternal death in our study. Notably, HIV/AIDS contributed about 3.5% between 2007 and 2012 but declined to 2.8% afterwards during the 2013–2019 period. The Prevention of Mother to child Transmission of HIV (PMTCT) program that included the provision of free care (including antiretroviral) and opt-out testing and counselling might have helped to reduce the contribution of HIV/AIDS to maternal mortality. Similar pattern of an initial increase and subsequent drop in HIV attributable maternal deaths was also observed in South Africa – a country with significant HIV burden [
30]. However, the PMTCT intervention strategies in Nigeria should be strengthened, and government should increase funding as the foreign donor interventions are dwindling.
Our study indicated a reduced contribution of Sickle cell disease to maternal death from 2.8 to 1.2%. This may be related to the fact that a dedicated unit caring for pregnant patients with sickle cell disease has emerged in our department [
31]. Thus, increasing experience and use of evidence-based protocol in the management of such patients might have been responsible for the reduced trends. However, public enlightenment campaign on premarital and prenatal diagnosis and counselling can help prevent the condition and its eventual maternal complications.
Relationship between socio-biological factors and leading causes of maternal deaths
Between 2007 and 2019, about half of the deceased were in the third decade of life. This is also similar to reports elsewhere in Nigeria and Sub-Saharan Africa [
4,
16,
19]. The death of women aged 30 to 40 years would have tremendous impact on family equilibrium and socio-economic productivity [
20]. In contrast, some studies reported that women within the ages of 20-30 years constituted most maternal deaths, possibly because of prevalent early marriage in such regions [
11,
14,
15].
Although grand multiparous women had a higher risk of maternal death from hypertension and haemorrhage, our study revealed that about 95% of maternal deaths occurred among women with parity less than 5. Thus, all women irrespective of parity should be protected from maternal death.
Strength and limitations
Although efforts were made during the monthly maternal mortality reviews that were conducted between 2007 and 2019 to clarify and update our entries based on the reviews, some information might have still been missed. One of the strengths of this study is that it is the first to utilise robust Joinpoint trends analysis tool to evaluate maternal mortality trends in Nigeria, to the best of our knowledge. The period of study was also relatively longer than obtains with most other published institutional data on maternal mortality. A longer period of trend analysis usually results in more valid conclusions. Furthermore, inferential statistics were used to improve the understanding of factors affecting maternal mortality. The conclusions from this study may be useful for designing interventions to reduce MMRs in tertiary hospitals that are in the urban cities in Nigeria and similar LMICs. However, the findings from our institutional study may not be generally applicable to the entire Nigerian population with different tiers of private and public health facilities and variations in regional and sociodemographic characteristics.