Hypoxic Patterns of Placental Injury: A Review

Histologic findings may indicate placental injury by decreased oxygen content, either occurring abruptly (acutely) or present chronically or as a fetal/placental reaction/adaptation to hypoxia, again developing relatively rapidly or during longer periods. Placental maturation is the most discriminative and by far the most important feature to assess in the diagnosis of chronic, in utero hypoxia, but it is also the most difficult to assess and the least-reproducible placental feature, even by experienced placental pathologists.¹⁵  It is impossible to assess placental maturation without a knowledge of gestational age, but, even if gestational age is known, the variation of maturation patterns is wide. The knowledge that, as pregnancy advances, terminal villi and syncytial knots become numerous, villous cytotrophoblasts and Hofbauer cells become less visible, the number of vascular profiles decrease, and the extracellular matrix of chorionic villi becomes denser, is of little help in individual cases, and the pathologist's experience is what matters. Thresholds of abnormality (Table 1) and reference range values for individual variables for gestational weeks can be helpful,^16–18  but they are not available for all hypoxia-related placental features.

A normal placenta shows heterogeneous maturation because better-oxygenized centers of cotyledons (placentones) are less mature (larger chorionic villi with less syncytial knots) than are their peripheral parts, where the less-oxygenized blood returns toward the uterus.¹⁹  The accelerated heterogeneous hypermaturity (Figure 1, A) is a sensitive feature of uteroplacental malperfusion.²⁰  Homogeneous placental maturation is always abnormal, the placentas being either hypomature (Figure 1, B) or hypermature (Figure 1, C). For obvious reasons, villous hypermaturity is easier to appreciate in preterm placentas and villous hypomaturity is easier in full-term placentas.

Villous hypervascularity (Figure 1, D; Table 1) can be diffuse or focal, both either fulfilling or not fulfilling the Altshuler rule of tens.^1,21,22  Capillary profiles can be highlighted by endothelial markers, for example, CD34, to facilitate counting; however, for the diagnosis of chorangiosis, the cutoff point of 20 lumens, instead of 10 lumens, on hematoxylin-eosin slide²³  should be used as more vascular lumens are seen by immunohistochemistry. Villous hypovascularity/avascularity (Figure 1, E) can be diffuse and associated, for example, with prolonged stillbirth, or focal, as in fetal thrombotic vasculopathy.^10,12,13 

The extracellular matrix of chorionic villi can be decreased, and the villous cores, thus, as pale as the intervillous space (Figure 1, F). Such a finding has to be distinguished from villous edema, where, in addition, a split artifact between the trophoblastic shell and the villous core is, at least focally, present (Figure 1, G).²⁴  An increased, extracellular matrix of chorionic villi, produced in excess by villous fibroblasts in response to intravillous hypoxia associated with intervillous hyperoxemia,^25,26  manifests as intense eosinophilia of villous cores (Figure 1, H).

Villous cytotrophoblasts are easily identified in the second trimester, but in the third trimester are not seen in abundance, except in chronic hypoxia that inhibits the trophoblast fusion and differentiation and stimulates proliferation, like in preeclampsia (Figure 1, I).^27–30  They can be more easily distinguished from endothelial cells and syncytiotrophoblasts using the E-cadherin/Ki-67 double immunostain, the former highlighting the cell membrane of villous cytotrophoblasts and the inner syncytiotrophoblastic cell membrane, and the latter staining the nuclei of proliferating cells (Figure 1, J).

Syncytial knots have transcriptionally inactive, condensed nuclei, without nucleoli, which can be either normal (granular, nonapoptotic) and due to tangential cutting (Figure 1, K), or smudgy (apoptotic) and due to nuclear aging³¹  (Figure 1, H). They should be distinguished from the transcriptionally active syncytial sprouting, which shows prominent syncytiotrophoblastic nucleoli (Figure 1, L) and is usually seen in the first and second trimesters,^9,32,33  and the clinical significance of which is not clear at this time. The number of knots is positively correlated with the length of time and the severity of the hypertensive diseases of pregnancy.³⁴ 

Villous Hofbauer cells synthesize proteins that stimulate a villous proliferation in response to hypoxia.³⁵  They are easily seen in the first and second trimesters, but not in the third trimester (Figure 1, M). A decrease in the amount of CD68⁺ cells during second and third trimester is continuous and gradual¹⁷  (Figure 1, N).

Stem obliterative endarteritis, a misnomer term applied to a swelling of intimal cells with occlusion of vascular lumen (Figure 1, O), due to hypoxia-stimulated medial vasoconstriction, is essentially a herniation of smooth muscle cytoplasm into the vascular lumen. It is linked to various pregnancy complications, particularly preeclampsia and fetal growth restriction (FGR),¹²  but also umbilical cord hypercoiling, where it tends to be associated with stem perivascular edema.³⁶ 

These developmental changes can be primary due to an intrinsic placental defect or secondary to intervillous hypoxemia. The patterns should not be confused with the clinical term placental insufficiency, which indicates a complication of pregnancy in which the placenta cannot carry enough oxygen and/or nutrients to the growing fetus and usually implies the presence of FGR.³⁷ 

The placental response to hypoxia does not follow a single pathway,²¹  the type of placental maturation and villous vascularity being the 2 key features used in classification of placental chronic-hypoxic patterns (Table 2). Features listed in Table 1 and discussed in the preceding section are also helpful.^21,31,38 

The preuterine pattern (PR) of chronic, hypoxic placental injury features, histologically, homogeneous placental maturation (ie, the chorionic villi are plump and of similar size from field to field); increased villous vascularity, including practically all cases of diffuse chorangiosis; decreased extracellular matrix of chorionic villi and villous cytotrophoblasts; increased Hofbauer cells; and increased nonapoptotic syncytial knotting^21,31  (Figure 2, A). Because of villous enlargement, the placenta looks hypomature, except for the focally increased syncytial knots, which is not, however, a constant feature. Placentomegaly and more-advanced gestational age are frequently seen.^21,39  The pattern is evoked by maternal hypoxemia secondary to decreased oxygen pressure in the environment (pregnancies at high altitudes), decreased oxygen binding capacity of the maternal blood (maternal anemia), air pollution and maternal smoking, increased distension of the uterus (multiple pregnancy), and maternal diabetes mellitus (abnormal oxygen-hemoglobin dissociation curve).^21,38–44  Clusters of multinucleate giant cells in the decidua basalis and excessive numbers of extravillous trophoblasts are less commonly seen in this than in other patterns of diffuse hypoxic placental injury,²¹  most likely because of the association of PR with deep trophoblastic, myometrial invasion, which was proven, at least in maternal anemia.⁴⁵  This pattern has a better prognosis than other types of chronic hypoxic injury,²¹  probably because of hypoxic preconditioning and resistance to ischemia-reperfusion injury during labor, as was proven in pregnancies at high altitudes⁴⁶  and multiple pregnancies.⁴³ 

The uterine (UH) pattern of chronic hypoxic placental injury (Figure 2, B) is typically seen in preeclampsia and late-onset FGR.^31,38  The term uterine is used here to refer to the maternal portions of the uteroplacental circulation, including its myometrial and decidual segments, which are the cause of pathologic placental changes. It is favored over uteroplacental to avoid confusion as to whether the maternal or the fetal side of the placenta is primarily involved.⁴⁷  Heterogeneous placental hypermaturity, with only focally increased villous vascularity, villous cytotrophoblast density, Hofbauer cells, and nonapoptotic syncytial knotting, and a decreased extracellular matrix of chorionic villi are the characteristic features. As with PR, the focal hypervascularity is an adaptive mechanism, reaching the level of chorangiosis in some cases, whereas, in other cases, the villous capillary profiles remain between 7 and 9 per chorionic villus (incipient or emerging chorangiosis).²²  The UH pattern is frequently associated with other placental features of uteroplacental malperfusion related to shallow, trophoblastic invasion, such as an accumulation of extravillous trophoblasts in the placental membranes (Figure 3, A), characteristically with membrane chorionic microcyst formation (Figure 3, B), and in the chorionic disc (Figure 3, C), again, typically, with microcyst formations (Figure 3, D), decidual clusters of multinucleate trophoblasts (Figure 3, E) and basal-plate myometrial fibers, and occult placenta accreta (Figure 3, F).^{27,38,48–54}  Excessive amounts of extravillous trophoblasts must be distinguished from massive perivillous fibrin deposition/maternal floor infarction,⁴⁹  which is associated with FGR, impaired neurologic development, recurrent fetal loss, and stillbirth,^9,12,55  but not necessarily with UH. Pathogenesis of the massive, perivillous fibrin deposition is different,⁹  but it can evoke fetal hypoxia by virtue of eliminating a substantial amount of functional placental parenchyma. I regard the above-presented extravillous trophoblasts lesions, along with the decidual arteriolopathy (both hypertrophic and atherosis), as the complementary criteria for this type of chronic placental hypoxia in histologically borderline cases.²¹  The UH and associated histologic findings cluster with severe preeclampsia, but not with mild preeclampsia; gestational hypertension; hemolysis, elevated liver enzymes, and low platelets (HELLP); or eclampsia⁵⁶  There is growing evidence that there are basic differences in the pathogenesis of mild and severe preeclampsia, the former usually occurring as a late onset, and the latter, usually having an early onset,^57–60  to which we recently contributed the molecular basis.⁶¹ 

The postuterine (PU) (postplacental) pattern of chronic hypoxic placental injury is due to primary villous changes resulting in decreased intake of oxygen from the intervillous space, as in retained stillbirth (Figure 2, C), subsets of FGR (Figure 2, D) and preeclampsia, and fetal thrombotic vasculopathy (only focally)^13,38  (Figure 2, E). Clinically, abnormal Doppler flow velocity waveforms obtained from umbilical arteries that reflect the downstream blood flow impedance may give an indirect evidence of vascular tree abnormalities.⁶²  The fetoplacental blood flow is compromised to a far greater extent in FGR with absent end-diastolic blood flow, such that maternal blood leaving the placenta has higher oxygen content than it does under normal circumstances.⁶³  The PU features homogeneous placental hypermaturity and hypovascularity, with slender, pencillike chorionic villi; so-called terminal villous hypoplasia¹³ ; increased extracellular matrix of chorionic villi and apoptotic syncytial knots; and decreased villous Hofbauer cells and villous cytotrophoblasts.³¹  The clinical, umbilical cord compromise with focal placental lesions of decreased fetal blood flow, which is frequently a random pregnancy accident, did not correlate with diffuse PU, although they can produce the focal PU, so-called stasis-induced thrombotic vasculopathy, with clusters of avascular, fibrotic, and occasionally hemosiderotic chorionic villi.^64,65  Also, severe, mass-forming, congenital anomalies that interfere with blood return from the placenta to the fetus can produce the placental stasis-induced thrombotic vasculopathy, but not the diffuse PU.⁶⁶ 

Hypomature placentas (placentas with maturation defect) are pale, normal sized or even larger than normal, with plumper terminal chorionic villi, superficially resembling intermediate villi,⁶⁷  and show defective formation of both terminal villous sinusoids (although numerically usually normal) and vasculosyncytial membranes (Figure 2, F), which can be better highlighted with CD34 immunostain (Figure 2, F inset). Although the vasculosyncytial membranes are normally poorly formed in early pregnancy, they are well formed in term pregnancies and are absent in only 5% of chorionic villi at term.¹²  In this type of placental pathology in pregnancies of 35 weeks or longer, on average, 1 vasculosyncytial membrane per terminal villus was found as compared with, on average, 3 per chorionic villi normally seen.⁶⁸  Villous hypomaturity is responsible for 22.5% of intrauterine deaths⁶⁹  and has a 70-fold increased risk of associated fetal death, with a 10-fold risk for recurrence, compared with baseline.⁷⁰  The fetuses die because of hypoxia,¹  but they can be rescued by earlier delivery.⁷⁰  There is an association between maternal diabetes mellitus⁷¹  and fetal anomalies⁷²  and maturation defect placentas, but vasculosyncytial membranes are also decreased in umbilical cord hypercoiling.⁷³  Diabetes mellitus is a state of chronic oxidative stress,⁷⁴  and glycemia appears to have an affect on the capillary, but not the stromal component, of chorionic villi.⁷⁵  Placental angiogenesis is stimulated by insulin via ephrin-B2 expression, a signaling molecule implicated in sprouting.⁷⁶  Therefore, this subtype of PR chronic hypoxic pattern appears to include the PU component because of fetal hyperinsulinemia.

The above-discussed patterns of chronic hypoxic placental injury are of a developmental origin because they are programmed and initiated early in pregnancy, some of them probably due to fetal or maternal genetic causes, and are usually fully developed at the turn of the second and the third trimester or later. Serum biomarkers of preeclampsia have been shown to be altered at as early as 7 weeks gestation, indicating that the onset of placental abnormalities in preeclampsia occurs even earlier than the onset of maternal blood flow, when failure of transformation of spiral arteries are thought to occur.⁷⁷  Maldevelopment of the maternal spiral arteries in the first trimester predisposes to placental dysfunction and suboptimal pregnancy outcomes in the second half of pregnancy.⁷⁸  The UH is, therefore, a maternal, hypoxemia-induced, focal-adaptive villous change, whereas PU is associated with intervillous hyperoxemia. The patterns may predispose a patient toward a greater vulnerability for superimposed, acute hypoxic lesions (overlap patterns/lesions, see below) or may sometimes protect against acute hypoxia in labor because PR can be an adaptive placental response and, in fact, protect the fetus against birth-related hypoxia.⁴⁶ 

Although characteristic, histologic features were originally described for the above patterns,³¹  I have proposed, based on the histologic picture, that one can predict what the pathogenesis of villous hypoxia is because the 3 patterns cluster with various clinical settings and associated placental findings.⁵⁶  The practical value of using the histologic terms PR, UH, and PU lies in keeping the pathologists aware that there is no single pattern of chronic hypoxic placental injury and, therefore, in expanding the pathologists' armamentarium in this respect.

In my experience,²¹  UH is most commonly associated with severe preeclampsia and the HELLP syndrome; PU, with abnormal Doppler results, induction of labor, clinical umbilical cord abnormalities, cesarean section rate, and FGR; and PR, with multiple pregnancies. Overall, PU is the most ominous and PR the least ominous histologic subtype of chronic in utero hypoxia. Sensitivity of placental injury from diffuse, chronic hypoxic placental injury is less than 50% for the main clinical conditions known to be associated with in utero hypoxia, and most cases of preeclampsia, pregnancy-induced hypertension, diabetes mellitus, nonreassuring fetal heart rate tracing, abnormal Doppler results (absent or reversed end-diastolic umbilical artery flow), and FGR were found in patients without diffuse placental hypoxic patterns. On the other hand, the high (>90%) specificity of the chronic hypoxic patterns of placental injury means that, in the presence of those histologic patterns, the clinical manifestations of chronic in utero hypoxia are very likely; therefore, the positive identification of one of the patterns is significant and unlikely to occur solely by chance.

Infarction is the tissue necrosis due to acute blood and oxygen deprivation either to placental membranes or decidua basalis (laminar necrosis) or villous tissue (villous infarction).

Membrane laminar necrosis (Figure 4, A), decidual, trophoblastic, or mixed, is a band of coagulative necrosis at the trophoblast/decidual interface and is associated with FGR,⁷⁹  maternal hypertensive disorders, and other conditions linked to in utero hypoxia.^48,80  It shows a characteristic evolutionary pattern (Figure 5, A). The significance of isolated amnion necrosis⁸¹  has not been independently confirmed, yet but most cases of amnion necrosis are meconium-induced (very common), whereas focal necrosis of the media of umbilical arteries of the umbilical cord is a rare response to meconium.^1,82  Laminar necrosis of the decidua basalis has been also described.⁸³  (Figure 4, B).

Acute restriction of focal, spiral artery blood flow or premature detachment of the placenta from the uterus causes villous infarction (Figure 4, C and D), the historically most frequently (or readily) diagnosed placental lesion, which starts to develop 2 to 4 hours to 4 days after an acute hypoxic event,¹³  the incidence thereof depending on gestational age, with preterm preeclamptic placentas being more vulnerable.^57,84  Early coagulation necrosis of the placenta underlying a retroplacental hematoma of placental abruption begins in 4 to 24 hours¹³  but is not complete after more than 24 hours.⁸⁵  A sequence of acute hypoxic placental lesions leads to formation of placental infarction, from congestion through agglutination of villi and intravillous stromal hemorrhage, at which time, it is irreversible (red infarction) (Figure 5, B). The accurate time frame of these sequential changes in humans is impossible to determine because, unlike, for example, in the myocardium, it is not possible to correlate the infarction histology with such factors as pain, serology, or electrocardiography, but results from monkey experimental studies, which revealed sequential placental infarction changes, were helpful.⁸⁶  Villous infarctions have their diagnostic limitations because they occur not infrequently in otherwise uncomplicated pregnancies, indicating a substantial placental reserve. In an attempt to increase the specificity of the diagnosis, only infarctions in a central/paracentral location and occupying more than 5% to 20% of the placental parenchyma, are regarded as diagnostically significant.^12,14,84  I use the lower threshold: central/paracentral infarction of at least 5% of placental parenchyma, and with that threshold, the sensitivity reaches 40% for preeclampsia.³⁸  The sensitivity can be higher if the noninfarcted placental parenchyma shows features of diffuse global hypoxia (an overlap lesion, see below).

Of fetal hypoxic lesions, mild erythroblastosis of fetal blood (Figure 6, A) is regarded as the only reliable (and, therefore, the best) evidence of chronic, in utero hypoxia; however, fetal erythroblasts can also be released from their stores in response to acute hypoxia.^9,13  Apart from hypoxia, nucleated red blood cells in fetal circulation can increase from infection; maternal diabetes mellitus; fetal anemia, as in erythroblastosis fetalis; and acute blood loss.¹⁰ 

Fetal erythroblastosis is stimulated by an increase in erythropoietin in response to hypoxia with a few hours' delay, with its umbilical blood level correlating with umbilical arterial pH.^87,88  The finding of appreciable amount of nucleated red blood cells in the placental circulation requires 6 to 12 hours or more, with umbilical blood, nucleated red blood cells being the gold standard.⁸⁹  However, a sample of umbilical blood may not be obtained at delivery, so nucleated red blood cell counts in histologic placental sections can be used as a surrogate test for fetal hypoxia.⁸⁹  Placental erythroblastosis may require severe, ongoing hypoxic stress of at least days', if not weeks', duration, and stillborn fetuses, dying of hyperacute etiologies, characteristically do not exhibit normoblastemia.²⁴ 

Some authors advice counting nucleated red blood cells per 100 white blood cells because the placental counts correlate well with umbilical blood counts^1,89 ; however, that is not practical, and most authors use subjective⁹  or semiquantitative⁷³  criteria to assess the number of nucleated red blood cells. I regard more than 1 erythroblast per high-power field in the third trimester of pregnancy as abnormal. Using this approach, nucleated red blood cells were found with increased frequency in association with acute and chronic, placental hypoxic lesions³⁸ ; multinucleated, trophoblastic giant cells in decidua⁵² ; and, clustered with PR.⁵⁶  Some authors found elevated nucleated red blood cells not only in primarily maternal or primarily chorionic disc pathology but also in umbilical cord abnormalities.^73,90 

Another fetal product commonly seen in placental sections is meconium, free or in macrophages. With free pigment/pigmented macrophages in membranes, I usually make my decision on hematoxylin-eosin slides to distinguish between meconium and hemosiderin because of granular refractibility of the latter, even if both pigments coexist. Presence of either pigment does not, however, appear to be strongly associated with acute, in utero hypoxia⁹¹  or with chronic placental separation,⁹²  respectively. Hemosiderin forms in tissue after 2 to 3 days after bleeding,⁶⁵  which can be useful when placental examination is a part of autopsy and timing is important. In that situation, I perform an iron stain if in doubt about nature of the pigment. Conspicuous meconium-laden macrophages at the chorionic surface of the placenta indicate that the fetus discharged meconium at least 2 to 3 hours before delivery. Abundant meconium-laden macrophages extensively deep in the placental membranes indicates meconium passage at least 6 to 12 hours before delivery.^82,93  Within 12 to 20 hours, the amnion can show signs of pseudostratification, cell degeneration, and even necrosis.⁹⁴  In the umbilical cord, necrotic arterial media with adjacent meconium-laden macrophages (Figure 6, B) indicates meconium discharge 12 to 16 hours before delivery by some authors⁹⁵  or at least 48 hours before delivery by other investigators. Meconium macrophages are diagnosed on histologic placental examination more frequently than meconium-stained amniotic fluid clinically,^21,48,38  possibly because after very recent meconium release, the membranes are not yet stained or because amniotic fluid can be cleared after a few days because of placental meconium phagocytosis. The primary cause of the meconium release (possibly hypoxia) may be difficult to separate from the toxic effects meconium has on fetal vessels. Histologic effects of meconium release depend not only on fetal stress/distress but also on meconium concentration in the amniotic fluid (thick meconium). Most authors agree that histologic meconium staining poorly correlates with clinical conditions that are potentially complicated by fetal hypoxia,^9,91,96  and histologic meconium is not a feature of fetal distress.⁹  It is more common in term pregnancies,⁵⁶  corresponding to the hypoxic-ischemic injury (also more common in term infants)¹³  and in association with chorionic disc diffuse and focal hypoxic lesions.^38,97  I share the opinion that the presence of meconium macrophages is a nonspecific feature of fetal hypoxia and indicative of fetal stress, rather than distress, in most cases,⁹¹  unless numerous meconium macrophages are seen deep in the decidua parietalis.

The term hypoxic overlap lesions means the coexistence of different hypoxic lesions or patterns from the same group,⁵⁶  such as, acute and chronic (Figure 7, A and B) or diffuse and focal (Figure 7, C through E). A decreased placental reserve by virtue of the presence of chronic, diffuse hypoxic injury can be more easily decompensated by an overlapping, acute hypoxia of a lesser degree (acute-on-chronic hypoxia) than would be required in an otherwise normal placenta.^13,38  Mild erythroblastosis of fetal blood was 3 times more common in placentas with other hypoxic lesions than it was in those placentas without them.³⁸  Therefore multiple placental hypoxic lesions, rather than as a single lesion, pose a greater risk of fetal growth restriction and neurologic impairment.^8,38,98 

In my collection of 5445 consecutive placentas older than 20 weeks, 47% of the placentas (n = 2564) did not show hypoxic features, and 53% (n = 2881) did. This illustrates the magnitude of the problem and the importance of diagnosing hypoxic lesions in the placenta. The purely placental hypoxic lesions dominated the purely fetal hypoxic lesions. The PR pattern was the most common, diffuse, chronic hypoxic pattern. More cases of preeclampsia, maternal diabetes mellitus, abnormal cardiotocography, abnormal dopplers, FGR, and cesarean sections and fewer cases of premature deliveries, congenital anomalies, and acute chorioamnionitis were seen in association with, or without, placental hypoxic lesions/patterns, respectively (Figure 8).⁹⁹ 

Because placental oxygen consumption is 4 times higher than fetal oxygen consumption is, the placenta is affected by in utero hypoxia first, followed, only later, by fetal compromise in some cases, as was proven in experimental studies for midgestation.¹⁰⁰  The placenta features a high reserve capacity so that placental signs of fetal hypoxia are far less common than are those of pure placental hypoxia.^21,38  Fetal hypoxia can also occur acutely (eg, after an acute cord compromise or placental abruption), without any associated histologic placental hypoxic lesions. On the other hand, because of the large reserve capacity, despite the presence of hypoxic lesions, the fetus can be normoxic and well, so an absolute correlation between placental hypoxic lesions and the maternal and fetal condition should not be expected, particularly because many complications of pregnancy occur either acutely or are not primarily due to placental pathology.

In summary, various types of hypoxic patterns have been discussed here, and lesions are depicted in Figure 9. The systematic identification of those patterns in a particular case should help to identify more placental hypoxic lesions, not just those limited to poor uteroplacental perfusion, and to clarify the etiopathogenesis of a significant proportion of the complications of pregnancy and abnormal fetal or neonatal outcomes. Because the placenta does not respond in a single way to hypoxia, the histologic changes should be explained in the clinical context and in conjunction with the gross placental and umbilical cord findings. Moreover, because the frequency and degree of expression of many of placental features are qualitatively and/or quantitatively gestational age-dependent, ^{17,18,57,101,102}  for precise evaluation (normal versus abnormal), the criteria presented in Table 1 may be insufficient, and comparison with tables of norms stratified by gestational age may be necessary, but those tables are not, however, available for all hypoxic variables. Such an approach can be helpful not only to better explain cases of perinatal morbidity/mortality and the mechanism of fetal injury, which could potentially be beneficial for management of future pregnancies, but also to help in medicolegal investigation of cases of perinatal morbidity/mortality.