Electronic databases and searching methods

A comprehensive search of five electronic databases–Ovid MEDLINE, EMBASE, CINAHL, ProQuest and Cochrane Library–was conducted on 28^th October 2022. The search strategy was created with the help of an experienced librarian (see S1 Text). The search was complemented by reviewing the reference lists of all included articles. Grey literature searches included unpublished studies identified through contacts with investigators who had conducted the studies. All recovered citations were uploaded into EndNote, and then screened using Covidence. Title and abstract screening, and subsequent full-text review was conducted by two independent reviewers, with conflicts resolved via discussion or consulting a third reviewer. Reports in languages other than English were first screened with an online translation tool (Google Translate) and subsequently translated by a native speaker to confirm eligibility and for data extraction and quality assessment.

Data extraction and quality assessment

Data for this review were extracted from included studies using Covidence. Variables of interest were country, sample collection year, types of outlets where samples were collected, country of manufacture, types of tests performed, total number of samples assayed, which pharmacopoeia or standard method was used and data relevant to outcomes of interest.

We assessed the quality of all included studies using the Medicine Quality Assessment Reporting Guidelines (MEDQUARG) checklist for reports of surveys of the quality of medicine [16]. Each study was graded on following domains: (1) survey details, (2) definition of substandard, falsified and degraded medicines, (3) types of outlets sampled, (4) sampling design and sample calculation, (5) type and dosage of units collected per outlet, (6) random sampling, (7) samplers’ detail, (8) packaging assessment, (9) statistical methods, (10) chemical analysis description, (11) method validation details and (12) blinding or chemical assessors. There is no established threshold using the MEDQUARG tool to define study quality, hence the studies were categorised using the approach of Torloni et al., which similarly evaluated studies on the quality of maternal medicines [13], where good methodological quality was defined as those with MEDQUARG score ≥ 6, and a score < 6 as low methodological quality. Two reviewers performed data extraction and quality assessment independently, with conflicts resolved via discussion or consultation with a third reviewer.

Data synthesis and statistical analyses

Data were presented descriptively. We reported on the prevalence of failed samples due to API content (whether too low or too high), sterility, pH and presence of impurities or related substances. Included studies and extracted data were organised by medicine type. We also reported the country where the study was conducted, and categorised that country income level, using the World Bank 2022 classification, and by type of outlets where the samples were collected.

Magnesium sulphate

Five studies assessed the quality of magnesium sulphate (Table 2), one from the USA (2014), two from Nigeria (2018 and 2021), one Ethiopian post-marketing surveillance report (2015) and one from the 2015 United Nations Commissions on Life-Saving Commodities (UNCoLSC) report that included 10 different countries- Burkina Faso, Kenya, Madagascar, Nepal, Nigeria, Tajikistan, Tanzania, Uganda, Vietnam and Zimbabwe [17–21]. The USA study found fungal contamination of Hamigera insecticola, Neosartrua hiratsukae, Penicillium chrysogenum, Penicillin rubens in a single bag of magnesium sulphate from a compounding pharmacy [17]. Two studies from Nigeria- 163 samples in the 2018 study and 90 samples in the 2021 study- reported that all samples passed sterility testing [19,21]. In the 2018 study, four of 163 samples did not meet API content specification and were reported as substandard [19]. In contrast, the 2021 Nigerian study [21] and post-marketing surveillance study from Ethiopia (13 samples) [18] reported all samples passed API content assays. The UNCoLSC study assessed a total 19 samples from 10 different countries in which two samples (one from Burkina Faso and one from Vietnam) failed to meet pH specifications [50]. Similarly, the Nigerian study (2021) reported 40 of 90 samples as poor-quality due to not meeting pH criteria [21]. Whereas the Nigerian (2018) and Ethiopian (2015) studies reported all samples passed the criteria for pH testing [18,19].

Aspirin

A total of eight aspirin quality studies were included (Table 3) [22–29]. All but one study (Iraq, 20 samples [28]) found evidence of substandard aspirin. Four studies, two studies- both from the USA conducted in 1979 [25,26], one from Brazil (2010) [23] and one from Yugoslavia (1994) [27], detected free salicylic acid above the acceptance range in 22 samples (12.8%), 44 samples (25.9%), five samples (62.5%) and 19 samples (55.9%), respectively. Another Brazilian study (2006) also detected free salicylic acid in aspirin samples, however the number of failed samples were not disclosed [24]. Only three studies from Nigeria (2012) [22], Cambodia (2004) [29] and the USA (1979) [25] assessed API content of aspirin. These found that API specifications were not met in two samples (20.0%), 16 samples (16.7%) and 44 samples, respectively. A Brazilian study (2010) found six samples (75%) failed to meet the dissolution criteria [23]. In the Cambodian study, physicochemical parameters such as hardness, weight variation, disintegration, friability and dissolution were assessed under quality evaluation, in which two samples (2.1%), one sample (1.0%), 16 samples (16.7%) and 32 samples (33.3%) failed in hardness, weight variation, disintegration and friability test, respectively [29].

Calcium supplements

Five studies assessed the quality of calcium supplements (see Table 4). All five studies tested for the presence of heavy metal impurities, such as lead or mercury [30–34]. Three studies from the USA (22 samples), Brazil (23 samples) and Pakistan (27 samples) identified lead contamination in 36.4%, 60.9% and 88.9% of samples respectively [32–34]. The other two studies, both from the USA in 1988 (ten samples) and 2005 (three samples), detected mercury and other trace elements in only a single sample in each study [30,31]. The Brazilian study also assessed API content and reported 31.1% (seven out of 23 samples) of calcium supplements were substandard with an API content below the specification range by Brazilian Pharmacopoeia [32].

Antihypertensive medicines

We found 15 studies on the quality of antihypertensives (Table 5) [35–49]. Seven studies evaluated the quality of amlodipine- two studies from India and Nepal found no evidence of poor-quality amlodipine [36,38]. In contrast, a Nigerian study reported three of ten (30.0%) amlodipine samples were counterfeit as they were unregistered within regulatory bodies, and their API content did not comply with British Pharmacopoeia standard.[37] A study in Cambodia found seven amlodipine samples (8.9%) did not meet the API content and uniformity criteria according to the US Pharmacopoeia and British Pharmacopoeia [40]. Another two studies—one conducted in 10 different African countries [35] and the other in Nigeria [41]- reported 87 (28.5%) and 48 (24.6%) amlodipine samples as substandard, respectively, due to failure to meet API specification. An additional Nigerian study reported four (50.0%) samples as substandard: one sample (12.5%) did not meet the API specification and three (37.5%) samples failed in friability and hardness criteria [39].

Two studies assessed nifedipine quality- one from Germany reported 60.3% samples of nifedipine did not comply with the German Federal Health Authority and US Pharmacopoeia API specification, and 27.9% of packaged and 76.5% of unpackaged nifedipine did not pass the light stability test [43]. Whereas, the other study from Nigeria concluded that 78 nifedipine samples (76.5%) were of poor quality, with 31 (30.4%) samples failing to comply with API specifications of US Pharmacopoeia and International Pharmacopoeia, and 76 (74.5%) samples showing nitrophenylpyridine impurities above the acceptance range [42].

Three studies from Nigeria [45], Rwanda [44] and Brazil [46], assessed methyldopa quality. A Nigerian study and Brazilian study reported one (25.0%) sample and ten (50.0%) samples as substandard, due to not meeting the API specification [45,46]. The study from Brazil also reported that ten (50.0%) samples failed to meet the criteria of API content. In contrast, the Rwandan study reported all three samples cleared the physicochemical parameters tested immediately after their purchase [44].

Three studies assessed enalapril quality- two from Brazil [46,47] and one from Turkey [48]. One of the Brazilian studies (2011; 20 samples) reported 100% of samples were substandard due to failure to meet Brazilian Pharmacopoeia API specification [46]. The study from Turkey reported that all eight samples met the criteria of API content and friability, however, all samples failed in the hardness test [48]. The other Brazilian study (2014) detected contaminants (including aluminium, gold, bromine, cerium, cobalt, hafnium, lanthanum, antimony, scandium, samarium, thorium, zinc, chromium) in all five samples [47]. Only one study on the quality of hydralazine was identified, conducted in the USA in which 12 (75.0%) samples were found to be poor-quality due to presence of impurities resulting from degradation i.e. hydrazine, above the acceptance limit [49].

Main findings

This is the first study to systematically evaluate the evidence on quality of medicines for HDP globally. We identified evidence of poor-quality of medicines for eight medicines—magnesium sulphate, aspirin, calcium supplements, and multiple antihypertensives, across high-, middle- and low-income countries. However, the majority of evidence come from studies in LMICs. The majority of studies (29/33) found evidence of poor-quality medicines used for the prevention and management of hypertensive disorders of pregnancy. The highest number of studies (15/33) assessed the quality of antihypertensives- almost all of which found evidence of poor-quality medicines, and the lowest number of studies assessed quality of calcium supplements and magnesium sulphate, five each.

Strengths and limitations

To the best of our knowledge, this systematic review is the first to assess the quality of multiple medicines used for HDP globally. More than half the studies we assessed were methodologically of good quality, suggesting robustness in our findings. We aimed to minimise the possibility of bias by using a broad and systematic approach to identify eligible studies without any limitations on language, geographical region and year of publication. In addition, we searched for information in the grey literature to reduce publication bias and maximise data capture. Despite these efforts, there may be additional reports that are not publicly available or were otherwise inaccessible. For example, some national authority bodies may be reluctant to release study results showing a high prevalence of poor-quality medicines.

Interpretation

In light of the examples of poor-quality HDP medicines that we identified—which crossed all medicine classes, diverse countries and multiple quality measures—we are concerned that this may be having a negative impact on maternal and newborn health. For example, a WHO review on medicine regulatory systems in 26 sub-Saharan African countries [51] reported that the presence of unregulated medicines in the market. Using such unauthorized medicines without quality assessment creates a profound risk for individual health by prolonging illness that leads to treatment failure, increasing morbidity and mortality. Failure to manage hypertension potentially increases the risk of developing preeclampsia/eclampsia potentially leading to premature delivery. This, in turn, impacts newborn development with long term effects from birth to adulthood including increased risk of developing hypertension, cardiac disease, metabolic disease and chronic renal disease [52]. The adverse consequences of poor-quality medicines for HDP specifically are difficult to predict. In low resource settings, there may be a number of barriers to high quality care, across the health systems. As such the adverse consequences of poor-quality medicines may initiate a flow on effect of adverse outcomes for women, as for example, they may lack access to appropriate care for HDPs, including access to routine antenatal care or diagnostic tools for preeclampsia, access to safe induction or caesarean delivery if required and high-quality neonatal cate if preterm delivery is needed to save the mother’s life. Thus, it is difficult to predict this flow on effect. Therefore, it is necessary to prevent, treat and manage HDP with high quality medicines. Insufficient API or degradation could result in inadequate dosing which could impair or reduce clinical benefit. For instance, poor-quality aspirin, calcium, antihypertensive medicines and magnesium sulphate [17–19,21–27,29–35,37,39–41,50] may mean that preeclampsia is more likely to develop, or worsen. It should be acknowledged, however, that the optimal (or minimum) dosing required for many obstetric medicines has not been established [53,54]. Contaminants—such as those reported for calcium, enalapril and magnesium sulphate [17,30–34,47]—are worrisome. In the absence of comprehensive toxicology evaluations, we can only speculate that these substances might jeopardise women’s and newborn’s health [55].

Poor-quality medicines can also reduce health workers’ confidence in the effectiveness of treatments, and otherwise impair public trust in the health system [11,14]. In addition, it affects public health due to poor health outcomes and also have adverse impacts due to additional health systems costs, wastage of resources, and increased out-of-pocket costs [11]. For instance, a study by Ozawa et al (2018) estimated the economic burden of substandard and falsified antimalarial and antibiotics in LMICs to range from $10 billion to $200 billion [10]. Currently, there are no cost-effectiveness studies on impact of poor-quality maternal medicines specifically, however, a study has shown the significant costs of caring for women with preeclampsia [56], which would be incurred if not effectively prevented or managed.

While this review was limited to medicines for HDP, quality issues affecting other medicines used in pregnant women have been identified. A 2020 systematic review on the quality of antenatal corticosteroids for preterm birth in LMIC [14] identified substandard dexamethasone samples due to inadequate API. A separate 2020 systematic review on the quality of multiple maternal medicines in LMIC identified 34 studies, with the highest failure rates reported for ergometrine, oxytocin and misoprostol [13]. A 2016 systematic review on the quality of oxytocin in LMICs identified poor quality of oxytocin samples due to insufficient API [15]. Combined with the evidence presented in the current study, there is clearly cause for concern that multiple medicines used in the management of leading causes of maternal and perinatal morbidity and mortality have quality issues.

Identifying precisely where and why quality issues arise within the manufacture and distribution pathways is key to developing strategies to improve medicine quality. However, most of the studies we identified did not assess important quality parameters, particularly physicochemical assays (dissolution, disintegration and friability assays). We also could not clearly identify the cause of low quality, as sample collection processes and locations were not described in most studies. It is thus difficult to draw conclusions whether poor-quality is attributable to manufacturing issues, storage conditions or a combination of both, based on currently available evidence. No studies reported information on how medicines had been stored in the sampling setting. Given that multiple studies reported on the presence of degradants or impurities, understanding storage conditions is particularly critical to addressing systematic causes of poor-quality medicines. Future research should comprehensively assess and report on all medicine quality parameters using reference methods, to facilitate a better understanding of the root causes of medicine quality issues. Studies should ideally follow WHO survey guidelines for post-marketing surveillance assessing medicines quality [57]. In addition, evaluating medicine quality from different outlets and points along the supply chain, and their current regulatory status, should also be performed. While it is acknowledged that resource capacity is limited, particularly in LMICs, we strongly encourage national regulatory authorities to implement robust pharmacovigilance and post-market surveillance programs to monitor and address poor-quality medicines, including for maternal health.