Severe maternal morbidity and near misses in tertiary hospitals, Kelantan, Malaysia: a cross-sectional study

Severe maternal morbidity includes a broader category of women who suffered complications related to pregnancy, delivery and puerperium that were not necessarily associated with critical illness such as blood transfusion. The WHO criteria not only consider clinical disorders (haemorrhagic, hypertensive and other systemic disorders) but also severe management indicators to indicate the severity and to enhance the identification of severe maternal morbidity [3]. In this study, approximately 2 % (18.3 per 1000 live births) of the 21,756 study population had severe maternal morbidity, which is comparable with the proportions recently reported elsewhere (0.8 to 17.5 %) [7, 8, 10‐15]. The primary causes of severe maternal morbidity were postpartum haemorrhage (53.2 %) and severe pre-eclampsia (26.1 %). Blood transfusion and ICU admission criteria represented 48.9 and 17.0 %, respectively, of women with severe maternal morbidity in this study.

The WHO has proposed 25 criteria based on the presence of organ and system dysfunctions (cardiovascular, respiratory, renal, coagulation, hepatic, neurologic and uterine dysfunctions) via clinical, laboratory- and management-based parameters for identification of maternal near miss [3]. Even using similar proposed criteria, the studies published to date have reported considerable differences in the proportions of maternal near miss. Most studies have reported a maternal near miss prevalence ranging from 0.4 to 3.3 % [7, 8, 10‐21]. Only one study in India reported an extremely high proportion of maternal near miss of 12.0 % based on a retrospective record review [22]. One large study, WHO Multicountry Survey on Maternal and Newborn Health (WHOMCS) reported the prevalence of maternal near miss of 0.8 % among 314,623 women attending 357 health facilities in 29 countries in Africa, Asia, Latin America and the Middle East (2538 maternal near miss and 486 maternal deaths) [23]. Twelve countries were involved in Asia, excluding Malaysia. Another recently published study in Nigeria also reported a prevalence that was within the reported range (1.6 %) [24].

The maternal near miss incidence ratio found in this study was 2.2 per 1000 live births (0.2 %) and was slightly low compared to other studies. As per the WHO criteria for maternal near miss, shock (8.5 %), severe hypoperfusion with lactate >5 mmol/L (21.3 %) and transfusion of five or more units of red blood cells (61.7 %) were the most common clinical, laboratory- and management-based parameters, respectively. Management-based parameters were the most frequently associated criteria for near miss (85.1 %) followed by laboratory-based criteria (40.4 %). This finding strengthens the specificity of management-based criteria in detecting severe obstetric cases for maternal near miss [25]. Postpartum haemorrhage, abnormal placental invasion and severe pre-eclampsia were the leading causes of maternal near miss (55.3, 21.3 and 14.9 %, respectively).

Considering the differences in healthcare systems and populations, the slightly lower prevalence of maternal near miss was justified. First, in this study, the women screened for organ and system dysfunction criteria for maternal near miss were from a pool of women initially classified with severe maternal morbidity. Although a slightly lower prevalence of maternal near miss was identified, the prevalence of severe maternal morbidity was within the reported range. Second, the data collection allowed clarification of doubts regarding the records obtained from the health personnel who took care of the women. This finding indicates that the data were real and unlikely due to underreporting. Steps were taken to ensure the data quality including pre-entry checking for completeness and double-checking medical records for unclear cases.

Third, ectopic pregnancy, which has a different aetiological pattern for maternal morbidity, accounted for approximately 8 % of maternal near miss cases similarly reported elsewhere [16, 21]. This figure is lower than that found in other studies (approximately 12 %) with high maternal near miss prevalence [13, 24]. Studies including a high number of early pregnancy losses in the numerator and using deliveries or live births as the denominator would inflate the prevalence of maternal near miss [26].

Maternal deaths are tragic events in obstetrics; however, severe maternal morbidity was estimated to be 100 times more common than death [27]. Our study found that severe maternal morbidity was almost 200 times more common than death, with almost 400 episodes affecting approximately 22,000 women in these two tertiary centres. Maternal death indeed constitutes the most obvious manifestation of severe morbidity related to pregnancy, childbirth and puerperium.

Sixteen maternal deaths occurred in Kelantan during the study period, seven of which were due to direct causes; three, indirect causes; and six, fortuitous deaths. Direct maternal deaths are due to obstetric complications, whereas indirect deaths are due to previously existing diseases or from a disease that develops and is aggravated during pregnancy. The 10 maternal deaths were caused by haemorrhage (three cases), eclampsia (one case), ectopic pregnancy (one case), amniotic fluid embolism (one case), infection (two cases) and undetermined (two cases). The WHO estimates that haemorrhage accounted for 27.1 % of maternal deaths; hypertensive disorders, 14.0 %; and sepsis, 10.7 %. The other deaths were due to abortion (7.9 %), embolism (3.2 %) and all other direct causes of death (9.6 %) worldwide [28].

In addition to prevalence, other indicators such as the maternal near miss mortality ratio and mortality index were used to describe the obstetric care. The maternal near miss mortality ratio estimates the complexity of care and refers to the ratio of maternal near miss cases and maternal death. This ratio also represents the proportion of maternal near miss cases that progressed to maternal death; the higher the ratio, the better the quality of care that the women received [4]. During the study period, two maternal deaths occurred and 47 cases of maternal near miss were identified in the two tertiary centres, revealing a maternal near miss mortality ratio of 23.5. Thus, for every 24 to 25 maternal near miss conditions, one maternal death occurred. The occurrence of at least four maternal near misses to every one maternal death also confirms other reports in the literature that justify the study of near miss cases [29].

Clearly, maternal near miss cases are more likely to die in resource-poor settings. For example, using similar WHO criteria [4], studies conducted in South Africa (Nigeria, Tanzania and Ghana) and Pakistan observed near miss to mortality ratios ranging from 1.5 to 2.5 [8, 17, 24, 30], while Brazil, India, Nepal and Iraq had near miss to mortality ratios ranging from 3.3 to 8.6 [7, 10‐13, 15, 16, 18‐20]. The ratio reported in this study is relatively high but is consistent with a study performed in China (23.0) [14] in which a similar method of data collection was applied. Higher ratios indicate a low maternal mortality ratio and better quality of care [4]. An overall increase in the ratio was observed in Brazilian studies with data collected from tertiary maternity hospitals from 2008 to 2012 [7, 10, 12, 16, 19, 20], except for a study by Lotufo et al. [15] that was conducted with data from a general intensive care unit. The increase in the near miss to mortality ratio over the period reflects an improvement in obstetric care. Therefore, instead of a single estimation, yearly estimations may facilitate monitoring and improving the quality of care provided.

The mortality index is an indicator to represents an estimate of performance. This index refers to the number of maternal deaths divided by the number of women with maternal near miss and maternal death and is expressed as a percentage [4]. Our findings revealed a mortality index of 4.1 %, which is comparable to the 4.2 % mortality index reported by Shen et al. (2013) in China. A high index (>20 %) indicates low quality obstetric care for severe cases, in which more women with severe conditions die. In contrast, a low index (< 5 %) indicates better quality of care, with fewer women with severe conditions dying [4]. Thus, the health facilities in this study are performing well regarding management of complex and severe cases.

High mortality indexes ranging from 22.9 to 40.8 % were reported in several studies in Africa, Latin America and Pakistan [8, 17, 19, 20, 24, 30]. Of these studies, two were the early retrospective studies in Brazil (2008 to 2009) [19, 20] and the highest index was reported in Nigeria [24]. In contrast, other studies in Brazil, Iraq, India and Nepal reported indexes between 10.4 and 18.1 % [7, 10‐13, 16, 18].

Haemorrhagic and hypertensive disorders were the two leading causes of severe maternal morbidity and maternal near miss in our study. This finding is consistent with other published studies in developing countries [11, 18, 30]. These obstetric events, which correlate with death, were also identified as the leading causes of maternal death in Malaysia [2].

Haemorrhagic disorders, primarily postpartum haemorrhage, ectopic pregnancy and abruptio placenta, account for approximately 50, 8 and 6 % of the severe maternal morbidity cases, respectively. A similar pattern observed in the maternal near miss cases indicated the potential progression of haemorrhagic complications to near miss and death. The finding that postpartum haemorrhage contributes the largest proportion is in line with the findings of other severe maternal morbidity and maternal near miss studies (36.1 to 48.5 %) [14, 18, 30]. Most of the cases of postpartum haemorrhage were due to uterine atony, consistent with studies reported elsewhere [15, 31, 32].

Ectopic pregnancy occurs in 1 to 2 % of all pregnancies and is the primary cause of morbidity and mortality in the first trimester of pregnancy [33]. The occurrence of ectopic pregnancy was 3.2 % [34] of severe maternal morbidity cases and approximately 12 % [15, 24] of maternal near miss cases. In our study, the occurrence of ectopic pregnancy was approximately 8 % in both conditions. The prevention of ectopic pregnancy in severe maternal morbidity cases remains a challenge as no risk factors have been identified that can predict severe intra-abdominal bleeding [35].

Abruptio placenta, the single most important cause of antepartum haemorrhage, complicates approximately 0.4 to 1 % of pregnancies [36], accounting for 7.1 % of severe maternal morbidity cases and 6.4 % of maternal near miss cases in our study, within the range reported elsewhere [15, 24, 37].

Hypertensive disorders, specifically pre-eclampsia, were the second highest cause of morbidity in our study. In spite of the high proportion of pre-eclampsia cases, the relatively low proportion of eclampsia cases may suggest adequate prevention of seizures. The progression to organ dysfunction occurred in seven cases (6.8 %) of pre-eclampsia and in two cases (9.1 %) of eclampsia. Appropriate and timely obstetrical care such as administration of magnesium sulphate and delivery of the placenta is crucial for preventing of morbidity and mortality. Even so, pre-eclampsia, which was thought to be a self-limited entity, appears to cause real damage to cardiovascular endothelial function [38]. Hypertensive disorders were also strongly associated with an increased risk obstetric complications such as abruptio placenta, thrombocytopenia and pulmonary oedema [39, 40].

Fluid overload imposed on severe pre-eclampsia patients may increase the risk of acute pulmonary oedema. In our study, acute pulmonary edema was prevalent in 2.5 % of severe maternal morbidity cases and in 6.4 % of maternal near miss cases. A much lower occurrence (0.9 %) of severe maternal morbidity cases was observed [15]. Limited data were available for comparison based on the WHO criteria. Among the severe hypertensive disorders patients, maternal near miss cases and maternal deaths were unsurprisingly high (16.6 %) but only 0.8 % severe maternal morbidity cases were observed [39].

The risk of uterine rupture is a concern for patients with previous caesarean section delivery. A case-control study on uterine rupture by the intended mode of delivery using the United Kingdom Obstetric Surveillance System (UKOSS) reported that the risk of uterine rupture with planned vaginal delivery in women with prior caesarean section was low, at approximately one in 500 deliveries [41]. A subsequent UKOSS study showed that the risk was more importantly associated with prostaglandin induction and oxytocin treatment during labour [42]. Therefore, the use of prostaglandin and oxytocin should be cautiously used during labour to improve patient safety.

In our study, previous caesarean section was found in 22.5 and 36.2 % of women with severe maternal morbidity and maternal near miss, respectively, contributing to the vulnerable status of this population. This result is consistent with findings from other studies regarding severe maternal morbidity (9.5 to 34.9 %) [7, 8, 11, 14] and maternal near miss (18.1 to 56.7 %) [8, 10, 11]. Previous or current caesarean section is often associated with increased likelihood of hysterectomy [43], consistent with our study, which found that among the 19 women with hysterectomy, 17 underwent caesarean section (14 emergency and 3 elective) and 12 (63.2 %) had previous caesarean section. This result suggests increased anticipation towards postpartum haemorrhage and peripartum hysterectomy for cases with previous and current caesarean sections. However, the presence of anaemia was unlikely to contribute to the morbid condition.

Haemorrhagic complications had great potential for progression into near miss, which is consistent with the high proportions of blood transfusion and hysterectomy in this study. Blood transfusion of five or more units of red blood cells indicates severe haemorrhage and a life-saving measure, whereas fewer units may indicate moderate anaemia to enhance postpartum recovery. A high proportion of large blood transfusion (61.7 %) was observed in our study. This result was much higher than proportion of blood transfusion (24.0 to 46.3 %) observed in maternal near miss cases in some studies [14, 15, 21]. Therefore, continuous adequate supplies and early access to blood products are crucial for successful rescue.

For patient safety, peripartum hysterectomy due to obstetric haemorrhage was commonly performed when other treatment modalities failed. In the same studies, in addition to receiving large blood transfusion, hysterectomy was performed [14, 15, 21]. All women that consented to hysterectomy (primarily for uterine atony, abnormal placental invasion or uterine rupture) in our study facilities survived. This finding highlights the need for an available surgical team to perform hysterectomy whenever required or an early referral to centres providing such services.

In our study, ICU admission with obstetric complications corresponded to 3.1 per 1000 live births. The most frequent obstetric morbidities transferred to ICU occurred following hysterectomy (17/19, 89.5 %) and haemorrhage due to placental implantation disorders (10/13, 76.9 %). The ICU transferals were for immediate post-operative surveillance and for intensive treatment purposes such as shock. The admission of obstetric cases to the ICU was very low, at 0.3 % (less than 0.5 %) and the admission of maternal near miss cases to the ICU was barely above the recommended standard, at 73.5 % (70 %) [4]. Approximately 20 to 50 % of women with maternal near miss failed to receive ICU care from events such as postpartum haemorrhage, placental implantation disorders and eclampsia, corroborating with the substantial proportion (46.8 %) of women experiencing two and more organ dysfunctions. This result suggests a shortage of ICU beds or the need to review the number of readily available ICU beds for obstetric patients.

Women presenting with obstetric complications may require a higher level of care. Those women with organ dysfunction would be more appropriately managed in the ICU to provide optimum care and to minimize the number of multiple organ failures [44]. Although an evidence-based triage system to assist clinicians regarding maternal utilization of intensive care services is lacking, haemorrhagic and hypertensive disorders were the two most common disorders in which admission into the ICU was deemed necessary [45].

Characterization of maternal near miss conditions according to organ system dysfunction is unquestionably feasible in tertiary hospitals in which procedures for monitoring are routinely performed without resource constraints [46]. By occurrence, the coagulation/haematologic dysfunction was the most common event followed by uterine and cardiovascular dysfunctions in our study. These events were manifested by large blood transfusion, hysterectomy and severe hypoperfusion criteria.

Cases that developed complications on arrival or within 12 h of arrival must be separated from those cases that developed complications in the hospital setting because the former would indicate failure in accessing tertiary centres and/or referral systems. Approximately half of women with severe maternal morbidity (46.3 %) and maternal near miss (55.3 %) developed complications within 12 h of admission, of which 68.3 and 61.5 % were referred cases, which may partly indicate an issue of delay in the referral system. Second, interestingly, the mortality index for in-hospital patients was 5.0 % compared to 3.4 % for referred cases. The higher mortality index suggests that the quality of care provided to in-hospital patients needs to be further reviewed concerning clinical management.

The use of oxytocin for the prevention and treatment of postpartum haemorrhage, of magnesium sulphate for the treatment of eclampsia, of prophylactic antibiotics for caesarean section and of parenteral antibiotics for the treatment of sepsis was covered almost 100 %. The coverage of recommended interventions below 95 % should be interpreted as an opportunity to improve care [4]. Good adherence to the measurable standards of the WHO guidelines suggests good quality of care in these facilities.

High coverage of essential interventions was observed in various countries participating in the WHOMCS; 18 % of women with severe maternal outcomes (maternal near miss and maternal death) did not receive the indicated interventions [23]. In poor-resource settings, a lack of information regarding organ dysfunction and an inadequate assessment of severity may contribute to the suboptimal implementation of essential interventions and clinical management. However, in our study, the process indicators of basic interventions seemed to be widely practiced and did not provide additional information regarding tertiary hospitals.

Coverage of essential interventions was suggested as the first important step in analysing the issues related to the quality of care. However, notably, application of these single elements of care in the provision of a comprehensive care system may not be adequate [23, 28]. For an instance, in relation to the care for postpartum haemorrhage, apart from the use of oxytocin for its prevention and treatment, other aspects of care such as shock management, an adequate donated blood supply and prompt surgical interventions are essential. Similarly, for the care of eclampsia, the role of magnesium sulphate is fundamental but severe hypertension management, pre-delivery stabilization and airway restoration are also vital. Although our study fared better with respect to coverage of essential interventions, further reduction of maternal mortality can be achieved by addressing these deficiencies.

The relevance of the essential interventions in further reducing maternal mortality has recently been questioned [47]. For example, in the WHOMCS, first, the risk of death did not increased [OR (95 % CI): 1.3 (0.81, 1.97], P = 0.330] in women with severe maternal outcomes that missed the opportunity to receive essential interventions compared to women without missed opportunities. Second, paradoxical performance regarding maternal mortality was observed. Coverage of all the five essential interventions was low in countries with low MMR and high in countries with high MMR. Although the differences between countries with low, moderate, high and very high MMR were not significant for three of the essential interventions, the coverage were significantly different for prophylactic antibiotics for caesarean section and parenteral antibiotics for sepsis. Thus, elements of care other than the essential interventions play an important role in the survival of women with morbid conditions [23].

According to the WHO process indicator, women presented at the hospital with obstructed labour or uterine rupture should undergo operations within three hours of admission. Failure to do so reflects substandard care and suggests an intra-hospital delay in the management of obstructed labour [4]. In our study, the cases of ruptured uterus occurred beyond the time limit. However, this process indicator is limited to cases of obstructed labour and uterine rupture presented at the hospital and not for hospitalized patients, suggesting the need for an improved indicator of ruptured uterus or new indicators related to prolonged or obstructed labour.

In our study, more than half of the pregnancies were terminated via emergency caesarean section, indicating the urgency of the obstetric conditions. Babies delivered to mothers with severe maternal morbidity and maternal near miss require neonatal ICU admission more frequently, as similarly reported by Lotufo et al. [15]. Half of the indications for admission to the neonatal ICU in our study were linked to preterm birth. Stillbirths due to severe maternal morbidity contributed to 5.1 % of all stillbirths, while, abruptio placenta accounted for 77.8 % of stillbirths in severe maternal morbidity cases. Among cases with abruptio placenta, the perinatal mortality was 32.1 %. In addition to the severity of abruptio placenta, foetal survival depends on the gestational age [36], and most of the stillbirths occurred prematurely, between 25 and 35 weeks.

Notably, the causes of severe maternal morbidity and maternal near miss were similar to the leading causes of maternal death. This result is relevant for identifying interventions that can be performed during antenatal care and delivery to reduce maternal death. Auditing of these cases demonstrated to be feasible despite the large number of cases [48]. In addition, the process of auditing would lead to increased experience in tackling the determinants of maternal morbidity [48] and understanding the real demands at each level of health care [3].

Cases of ruptured uterus occurring in hospitalized women are an important indication of the seriousness of substandard care. The current process indicator has outlined that women presented at the hospital with obstructed labour or uterine rupture should undergo operation within three hours of admission. Modification and improvement of this indicator is highly required. This study suggests for the monitoring to begin from the time these diagnoses were made for hospitalized patients but with shorter time limit.

Of 10 maternal deaths, two deaths occurred in tertiary centres and four deaths occurred in hospitals with and without obstetrics and gynaecology specialists (the remaining four women were brought in dead). This result implies that, apart from patient factors, the occurrence of severe maternal morbidity and maternal near miss, similar to maternal deaths, may be influenced by the health care level. Research involving different levels of health facilities should be repeated to objectively identify issues related to pregnancy complications upon which appropriate interventions could be adopted. Apart from allowing comparisons between health facilities, a single health facility could monitor its progress over time.

The findings of this study should not be regarded as representative of Kelantan but indicative of a large hospital-based tertiary centre study. The maternal outcomes in community-based birthing centres and district level hospitals were not assessed. Utilization of the WHO maternal near miss criteria seems to be limited in smaller health facilities. The laboratory-based (for example, pH, PaO₂ and lactate) and management-based markers (for example, vasoactive drug management and hysterectomy) are less likely to be applicable in these health facilities. Therefore, we support severe maternal morbidity as the initial classification in the continuum that begins with any occurrence of complications of pregnancy, delivery and puerperium until death. It is proposed that this pragmatic criteria approach related to severe maternal morbidity be applied at poor-resource health facilities and the strict near miss criteria approach related to organ dysfunctions be applied for quantitative assessments and international comparisons. Thus, the strict near miss criteria should be incorporated as much as possible [49]. These approaches should complement each other to obtain information to improve maternal health.