Maternal and Fetal Outcomes of Pregnancies in Women with Atypical Hemolytic Uremic Syndrome

Abstract

Atypical HUS (aHUS) is a disorder most commonly caused by inherited defects of the alternative pathway of complement, or the proteins that regulate this pathway, and life-threatening episodes of aHUS can be provoked by pregnancy. We retrospectively and prospectively investigated 27 maternal and fetal pregnancy outcomes in 14 women with aHUS from the Vienna Thrombotic Microangiopathy Cohort. Seven pregnancies (26%) were complicated by pregnancy-associated aHUS (p-aHUS), of which three appeared to be provoked by infection, bleeding, and curettage, and three individuals were considered to have preeclampsia/HELLP syndrome before the definitive diagnosis of p-aHUS was made. Mutations in genes that encode the complement alternative pathway proteins or the molecules that regulate this pathway were detected in 71% of the women, with no relationship to pregnancy outcome. Twenty-one pregnancies (78%) resulted in a live birth, two preterm infants were stillborn, and four pregnancies resulted in early spontaneous abortions. Although short-term renal outcome was good in most women, long-term renal outcome was poor; among the 14 women, four had CKD stage 1–4, five had received a renal allograft, and three were dialysis-dependent at study end. We prospectively followed nine pregnancies of four women and treated six of these pregnancies with prophylactic plasma infusions (one pregnancy resulted in p-aHUS, one intrauterine fetal death occurred, and seven pregancies were uneventful). Our study emphasizes the frequency of successful pregnancies in women with aHUS. Close monitoring of such pregnancies for episodes of thrombotic microangiopathy is essential but, the best strategy to prevent these episodes remains unclear.

Pregnancy presents a challenge for the maternal immune system because of the need to protect the fetus from pathogens while preventing alloimmune injury by facilitating tolerance to paternal antigens. The challenge is focused on the feto-maternal interface in the placenta and requires a precise control of adaptive and innate immunity, including the complement system as exemplified by women with atypical hemolytic uremic syndrome (aHUS).

aHUS is a disorder most commonly caused by inherited defects of the alternative pathway of complement, or of the proteins that regulate it, and pregnancy can provoke life-threatening episodes.^1,2 Pregnancy-related aHUS (p-aHUS) affects one in 25,000 pregnancies in the general population,³ which contrasts with an incidence of 20% in women with preexisting aHUS in whom it causes significant maternal mortality and morbidity.⁴ Classically, p-aHUS presents in the postpartum period and was believed to have little effect on fetal outcome. However, preeclampsia, eclampsia, and HELLP syndrome (hemolysis–elevated liver enzymes–low platelet) are all strongly linked to complement dysregulation and consequently to aHUS that has been complicated by worse fetal outcomes.⁵

Despite the risks, current consensus proposes that each pregnancy should be considered individually and carefully planned.⁴ However, the knowledge base is small, and it needs to be expanded before women with aHUS can be given robust advice about pregnancy.⁶ Here we present data on pregnancies from well characterized patients with aHUS, recruited into the Vienna Thrombotic Microangiopathy (TMA) cohort.

In total, we document 27 pregnancies in 14 women, including nine pregnancies in four women that were followed prospectively.

Results

Clinical Aspects of p-aHUS

By the end of June of 2015, a total of 92 patients had been entered into the Vienna TMA Cohort. These included 30 (33%) with aHUS, of whom 20 were female; one with Shiga toxin–producing Escherichia coli (STEC)–HUS; and 61 with other causes of TMA (Figure 1). Fourteen of the 20 women with aHUS had been pregnant, and together they had had 27 pregnancies, of which nine (in four women) were followed prospectively and are described in detail in the Supplemental Data.

Figure 1.

Pregnancies among patients with aHUS. Flowchart of the Vienna TMA Cohort. *Developed p-aHUS in subsequent pregnancies.

The 14 women with aHUS who had pregnancies were aged 29±12 years (mean±SD) at diagnosis (Table 1). They segregated into three groups, depending on the timing of pregnancy in relation to diagnosis (Figure 1). Group A (14%) consisted of two individuals (A.1 and A.2) in whom aHUS was diagnosed before their first pregnancies, which occurred 6 and 14 years later. Group B (36%) contained five individuals (B.1 to B.6) who had had at least one uncomplicated pregnancy before their first episode of aHUS, that occurred a median of 6 years (range, 2–23 years) after their last pregnancy. Group C (50%) comprised seven women (C.1 to C.7) who presented with aHUS during or immediately after pregnancy. In five of these women, p-aHUS occurred during their first pregnancy, in the second trimester in two, the third trimester in two, and in the puerperium in one. Notably, patient C.6 initially presented with aHUS 1 year after the birth of her first baby and, after an uneventful second pregnancy, subsequently developed another episode during her third pregnancy. Patient C.7 presented with p-aHUS in the puerperium of her second pregnancy.

Table 1.

Manifestation of aHUS, treatment, and renal function during pregnancies in 14 women

Patient	Onset of aHUS		Preg	Prior First Preg CKD Stage	Adverse Events		Maintenance Treatment		As of 2015 (End of Study)
	Age	Treatment			During Preg	After Preg	Prior First Preg	After Preg	Age	CKD Stage	Mainten. Treatment
A.1	3	supportive, PEX	2	CKD-G1	1: none, with PI	1: none	none	1: none	24	CKD-G1	none
					2: proteinuria, hypertension, with PI	2: persistent proteinuria		2: none
A.2	24	none	3	CKD-G1T	1: none (KTX)	1: none	none	1: none	33	CKD-G2T	none
					2: none (KTX)	2: none		2: none
					3: none (IUFD gw 31)	3: none		3: none
B.1	28	PEX, ecul	1	unknown	1: none	1: none	none	1: none	31	CKD-G4	none
B.2	39	none	2	unknown	1: none	1: none	none	1: none	43	CKD-G3bT	ecul
					2: hypertension	2: none
B.3	37	PEX	2	unknown	1: none	1: none	none	1: none	41	normal	none
					2: none	2: none
B.4	34	PEX	1	unknown	1: none	1: none	none	1: none	42	CKD-G4T	ecul
B.5	57	none	2	unknown	1: none	1: none	none	1: none	60	CKD-G2	none
					2: none	2: none
C.1	20	PEX	3	normal	1: none	1: p-aHUS	none	1: none	23	CKD-G1	none
					2: none, with PI	2: none		2: none
					3: none, with PI	3: none		3: none
C.2	31	none	3	CKD-G3b	1: increasing creatinine, preeclampsia, fetal death	1: p-aHUS	none	1: none	53	CKD-G5D	none
					2: none, spont. abortion	2: none		2: none
					3: none, spont. abortion	3: none		3: none
C.3	28	none	1	CKD-G3b/4T	1: preeclampsia gw 26	1: p-aHUS	none	1: none	42	CKD-G5D	none
C.4	24	PEX	1	normal	1: spont. abortion	1: p-aHUS	none	1: none	27	normal	none
C.5	33	none	1	normal	1: proteinuria, HELLP	1: p-aHUS	none	1: none	42	CKD-G5D	none
C.6	19	PEX	3	normal	1: none	1: aHUS (+12 mo)	none	1: PI	32	CKD-G4T	ecul
					2: none, with PI (KTX)	2: none		2: PI
					3: none, with PI (KTX)	3: p-aHUS (+1 d)		3: ecul
C.7	27	PEX	2	unknown	1: none	1: none	none	1: none	33	CKD-G1T	ecul
					2: none	2: p-aHUS (+6 d)		2: PEX

Preg, pregnancy; mainten., maintenance; PEX, plasma exchange; PI, plasma infusion; KTX, kidney transplant; IUFD, intrauterine fetal death; gw, gestational week; ecul, eculizumab; spont., spontaneous.

Overall, seven pregnancies (26%) were complicated by episodes of TMA (p-aHUS), whereas 20 (74%) were not. Accordingly, the incidence of p-aHUS was 6.94 episodes per 1000 weeks at risk (95% confidence interval [95% CI], 2.79 to 14.27). In three patients, episodes of TMA appeared to be provoked by recognized triggers: infection in patient C.6; bleeding in C.1; and curettage after a spontaneous abortion in patient C.4. In addition, three women (C.3, C.4, and C.5) were considered to have preeclampsia or HELLP syndrome before the definitive diagnosis of p-aHUS was made. One individual (C.7) had an induced preterm delivery, but we have no information about why this was considered necessary.

Complement Gene Mutations

Mutations in genes that encode complement alternative pathway proteins, or the molecules that regulate this pathway, were detected in 10 of 14 (71%) in our p-aHUS cohort, a frequency within the range of 50%–80% reported for aHUS more generally.^7,8 Details of the mutations are shown in Table 2. On the basis of in silico analysis, functional studies, or previous association with aHUS, three of the patients (A.2, B.5, C.1) have mutations that are recognized as highly likely to be pathogenic and a further two (A.1, C.4) are recognized as probably pathogenic. The remaining five patients had rare variants that are also likely to be pathogenic.

Table 2.

Results of genetic investigations in 14 women with aHUS

Patient	Gene	Alteration in Protein	Domain	In Silico Predictions on Possible Functional Effect	Minor Allele Frequencya	Functional Effect	Risk Haplotype	Copy Number of CFH, CFHR1,-R2, -R3, and -R5
A.1	CFI	p.G342E	serine protease	probably damaging,1 damaging,2 deleterious,3 disease causing4	0%	not characterized42	het CFH H3	no alteration
	CD46	p.D257Vfsa42	SCR4	disease causing4	0%	FS causing, damaging43
A.2	CFI	p.I416L	serine protease	benign,1 tolerated,2 neutral,3 disease causing4	0.01%–1.29%	quantitative Factor I Deficiency42,44–47	none	het dele CFHR1,-R3
B.1	none		na	na	na	na	het CFH c.-331C>T	het dele CFHR1,-R3
B.2	none		na	na	na	na	hom CFH H3	no alteration
B.3	CD46	p.E234K	SCR4	benign,1 tolerated,2 neutral,3 poly4	0%	not characterized7	het CFH H3	no alteration
B.4	C3	p.D61N	MG1	probably damaging,1 tolerated,2 neutral,3 poly4	0%	not characterizedb	het CFH H3	no alteration
B.5	CD46	p.A353V	TM	benign,1 tolerated,2 neutral,3 poly4	0.29%–1.73%	deficient cell surface control of AP48–52	none	no alteration
C.1	C3	p.K104E	MG1	benign,1 tolerated,2 neutral,3 poly4	0%	not characterized	het CFH H3	no alteration
	C3	p.D1457H	MG8	probably damaging,1 damaging,2 deleterious,3 poly4	0.34%	not characterized
C.2	THBD	p.E560Q		probably damaging,1 tolerated,2 neutral,3 disease causing4	0.01%–0.02%	not characterized	het CFH H3	no alteration
C.3	none		na	na	na	na	het CFH H3/H8	na
C.4	CFHR5	p.E163Rfsa34	SCR3	disease causing4	0.09%–0.30%	FS causing, damaging53,54	none	no alteration
C.5	none	None	na	na	na	na	hom CFH H3	no alteration
C.6	CFH	p.N516K	SCR9	probably damaging,1 tolerated,2 deleterious,3 poly4	0.02%–0.04%	not characterized42,44,48,55	het CFH H3 hom p.E936D	no alteration
C.7	CFH	p.D748Nfsa10	SCR13	disease causing4	0%	FS causing, damaging	het CFH H3
	CD46	p.A353V	TM	benign,1 tolerated,2 neutral,3 poly4	0.29%–1.73%	deficient cell surface control of AP48–52		no alteration

CFH, complement factor H; CFHR1–5, CFH related protein 1–5; CFI, complement factor I; het, heterogeneous; CD46, cluster of differentiation 46; SCR, short consensus repeat; FS, frameshift; dele, deletion; na, not applicable; hom, homogenous; poly, polymorphism; MG, macroglobulin; TM, transmembrane; AP, alternative pathway; THBD, thrombomodulin.

^aMinor allele frequency was included on the basis of the Exome Sequencing (http://evs.gs.washington.edu/EVS/) and 1000 Genomes Projects (reported by dbSNP https://www.ncbi.nlm.nih.gov/SNP/); for patient C.3 only CFH was analyzed and CFH c.-331C>T was not included (CFH-H3/H8); CFHR5 was sequenced only in patients C.4 and C.5.

^bp.K65Q at a near site affects the C3 binding to factor H and membrane cofactor protein.

The genes CFH, CFI, and CD46 are inherited together with other regulators of complement activation as part of a gene cluster on chromosome 1. Penetrance of pathogenic CFH, CFI, and CD46 mutations for aHUS is incomplete but is increased when they are inherited on a particular haplotype (CFH-H3).⁷ Ten of the 14 patients in the p-aHUS cohort inherited the CFH-H3 risk haplotype that included a complement mutation in seven individuals. Notably, the two patients without complement mutations were the only ones that were homozygous for CFH-H3 (Table 2).

Specific mutations had no discernable effect on the timing of p-aHUS in relation to pregnancy. Thus, recognized disease-causing mutations were present in both women who developed aHUS years before they first became pregnant (A.1 and A.2); in at least three of the five women who had uncomplicated pregnancies before developing clinically severe aHUS (group B); and in five of the seven women in group C who presented with p-aHUS.

Pregnancy in Women with aHUS

We analyzed 18 pregnancies retrospectively (patients B.1 to B.5 and C.1 to C.7), none of which were managed with prophylactic infusions of fresh frozen plasma (FFP) or other preventative therapy. Six of these pregnancies (33%) were complicated by p-aHUS (Table 1).

None of the eight pregnancies in individuals in group B were complicated by p-aHUS. However, patient B.2 developed hypertension without proteinuria (Table 1). During their first pregnancies, patients C.2 and C.3 (who had received a renal allograft) were both diagnosed clinically as having preeclampsia, but proved to have TMA on renal biopsy (p-aHUS), and subsequently progressed to ESRD. Patient C.4 developed p-aHUS with ARF triggered by curettage after spontaneous abortion. She was treated with eight therapeutic plasma exchanges, and her renal function recovered completely. Patient C.5 developed typical clinical signs of p-aHUS after a stillbirth and an initial diagnosis of HELLP syndrome; no specific treatment was given, and she developed progressive CKD. Finally, patient C.7 presented with p-aHUS after an uneventful second pregnancy and developed ESRD despite therapeutic plasma exchange.

Nine pregnancies were followed prospectively (A.1, A.2, C.1, and C.6), one of which led to p-aHUS (C.6). Patient A.1 was originally diagnosed as having STEC-HUS when she presented with TMA at the age of three. She subsequently had 22 further episodes before a definitive diagnosis of aHUS was made when she was shown to have inherited a CFI and CD46 mutation on the CFH-H3 susceptibility haplotype. She has had two pregnancies, both treated with prophylactic FFP (3 ml/kg twice monthly; Table 3). The first pregnancy was uncomplicated, but she developed hypertension and significant proteinuria during the second, although her serum creatinine concentration remained stable (Table 4). There were no signs of complement activation during either pregnancy.

Table 3.

Details about aHUS treatment of four women during pregnancies and postpartum period

Patient	Preg	Year	Prospectively Followed	Maintenance Plasma Therapy before Pregnancy	Plasma Therapy during Pregnancy	Start (Gestational Week)	Stop (Postpartum Week)	TMA Episode
A.1	1	2009	yes	none	3 ml/kg twice monthly	15	continued	no
	2	2012	yes	none	3 ml/kg twice monthly	13	continued	no
A.2	1	2010	yes	none	2.5 ml/kg twice monthly	14	continued	no
	2	2011	yes	none	none	—	—	no
	3	2014	yes	none	none	—	—	no
C.1	1	2011a	no	none	none	—	—	yes
	2	2014	yes	none	10 ml/kg twice monthly	21	14	no
	3	2015	yes	none	10 ml/kg twice monthly	15	24	no
C.6	1	2002a	no	none	none	—	—	nob
	2	2013	yes	25 ml/kg monthly	25 ml/kg twice monthly	14	19	no
	3	2015	yes	25 ml/kg monthly	25 ml/kg monthly	—	1c	yes

Preg, pregnancy.

^aPregnancy before follow-up at our department.

^bPresented with acute and chronic TMA in renal biopsy 12 mo after delivery.

^cSwitch to eculizumab.

Table 4.

Renal function, BP, and signs of microangiopathic hemolytic anemia during 11 pregnancies

Patient	Preg	Gestational/pp. weeka				Kidney Function					Signs of Microangiopathic Hemolytic Anemia					Hypertension
A.1	1	13	28	36	3	S-Crea	0.67	0.69	0.69	0.93	Hb	11.1	10.4	10.1	11.7	BP	125/75	120/80	125/80	127/75
						UPE	386	320	397	488	PLC	234	182	170	299	Med	none
											LDH	143	171	165	164
	2	14	28	36	3	S-Crea	0.71	0.66	0.71	0.85	Hb	11.8	11.5	10.6	11.8	BP	130/70	130/80	130/80	130/80
						UPE	776	727	2208	1178	PLC	237	254	225	317	Med	none	N	N	N
											LDH	143	202	235	237
A.2	1	13	28	36	5	S-Crea	0.95	0.91	1.06	1.16	Hb	10.1	10.2	10.6	12.1	BP	125/88	125/79	147/105	120/80
						UPE	56	112	155	91	PLC	263	172	137	182	Med	M	M	M	none
											LDH	180	157	207	164
	2	14	27	39	2	S-Crea	0.86	0.85	0.85	1.15	Hb	10.5	10.6	11.2	10.9	BP	120/80	120/80	140/90	140/100
						UPE	47	164	280	65	PLC	308	254	194	332	Med	N	N	N, M	N, M
											LDH	169	143	215	215
	3	12	28	33	1	S-Crea	0.71	0.8	0.78	0.92	Hb	10.4	10.7	11.9	9.3	BP	125/75	123/84	123/84	130/89
						UPE	104	127	145	102	PLC	349	233	258	304	Med	N, M	N, M	N, M	N, M
											LDH	171	154	191	209
C.1	1	8	25	40	1	S-Crea	na		0.5	2.2	Hb	14.1	11.0	9.2	8.0	BP	115/75	110/70	na	PRES
						UPE	neg	neg	na	362	PLC	na	na	254	52	Med	none			na
											LDH	na		188	1964
	2	20	27	37	5	S-Crea	0.76	0.74	0.72	0.9	Hb	12.3	12.1	9.8	11.2	BP	110/90	125/85	125/78	125/78
						UPE	110	169	253	390	PLC	235	240	240	368	Med	M	M	M	none
											LDH	184	164	268	194
	3	13	27	37	3	S-Crea	0.74	0.71	0.79	0.89	Hb	11.3	8.9	12.5	11.2	BP	138/102	123/65	126/81	121/71
						UPE	373	342	456	290	PLC	333	284	222	335	Med	M	M	M	M
											LDH	185	205	191	na
C.6	1	8	23	36	na	S-Crea	na				Hb	14.1	12.3	na	na	BP	110/70	110/70	120/80	na
						UPE	neg	neg	neg	na	PLC	272	243	na	na	Med	none
											LDH	na
	2	14	28	38	3	S-Crea	1.46	1.71	2.01	1.96	Hb	10.0	8.4	8.9	9.3	BP	130/85	130/88	132/83	140/93
						UPE	101	85	202	366	PLC	279	236	191	398	Med	U, H, N	U, L, N	U, L, N	U, Ca, A
											LDH	154	na	184	215
	3	13	29	37	1	S-Crea	1.66	2.06	2.42	7.53	Hb	9.9	8.5	9.6	6.8	BP	155/95	133/71	135/90	158/82
						UPE	212	244	2848	7653	PLC	263	177	178	108	Med	U, L, N	U, L, N	U, L, N	U, Ca, A
											LDH	146	182	246	592

Preg, pregnancy; pp., postpartum; S-Crea, serum creatinine (mg/dl); Hb, hemoglobin (mg/dl); BP, BP (mmHg); UPE, urinary protein excretion (mg/g creatinine); PLC, platelet count (G/L); Med, medication; LDH, lactate dehydrogenase (U/L); N, Nifedipin; M, Methyldopa; na, not available; PRES, posterior reversible encephalopathy syndrome; neg, urine dipstick negative; U, urapidil; H, hydrochlorothiazide; L, labetolol; Ca, carvedilol; A, amlodipine;

^aColumns 1–3 refer to gestational week; column 4 refers to postpartum week.

After repeated episodes of TMA according to renal biopsy, patient A.2, aged 24, presented with ESRD and later successfully received a renal allograft. After this she had three pregnancies all of which were followed prospectively. The first was managed with prophylactic FFP (2.5 ml/kg FFP twice monthly) and was uncomplicated. Prophylactic FFP was not given during either of the subsequent pregnancies: the second was completely uncomplicated, but the third resulted in a late intrauterine death, probably caused by in utero infection. There was no evidence of complement activation during any of the three pregnancies.

Patient C.1 has compound C3 mutations and has inherited the CFH-H3 susceptibility haplotype. Originally, she presented elsewhere with p-aHUS after her first pregnancy and was treated with 26 plasma exchanges over 41 weeks, but did not require maintenance therapy after she recovered. Her two subsequent pregnancies were managed by us with twice monthly prophylactic infusions of FFP (10 ml/kg), during the pregnancy and into the puerperium (Table 3). Neither pregnancy was complicated by p-aHUS, despite an infection with influenza H1N1—a known trigger for aHUS—during the third pregnancy (Table 4). However, she did have decreased serum C3c concentrations for at least 6 months postpartum, despite normal C3 concentrations and otherwise unremarkable laboratory values.

Patient C.6 has a CFH mutation on a CFH-H3 haplotype and developed aHUS complicated by renal failure 12 months after an apparently uncomplicated pregnancy. She received a renal allograft and was treated thereafter with monthly prophylactic infusions of FFP (25 ml/kg monthly). She had two further pregnancies. In the first, we increased the prophylactic FFP infusions from once to twice a month from the first trimester until the 19th postpartum week (Table 3), and supplemented this with 1600 ml FFP infusions on the day of her caesarian section and the day after (Table 4). She continued her maintenance prophylactic treatment with FFP during her third pregnancy and was again infused with 1600 ml FFP before and after her caesarean section. Despite this, she developed TMA within hours after surgery, possibly provoked by infection. The hemolysis and thrombocytopenia resolved rapidly with FFP infusions, but she developed ARF in the kidney transplant and a renal biopsy confirmed acute TMA (Figure 2). FFP alone has been reported to be ineffective in this situation,^9–12 and so eculizumab was added, and this was followed by a steady improvement in serum creatinine. In retrospect, plasma haptoglobin had become undetectable around 4 weeks before the caesarian section, which might be suggestive of low levels of TMA despite the absence of symptoms or other evidence of microangiopathy (Table 4).

Figure 2.

Renal TMA in a patient with aHUS. Biopsy specimen of renal allograft of patient C.6 after third pregnancy showing (A) arteriolar TMA (acid fuchsin orange G stain, 200× magnification) and (B) glomerular TMA (Periodic Acid-Schiff stain, 200× magnification). In comparison, a renal biopsy after her second pregnancy showed (C) a histologically normal glomerulum with interstitial fibrosis (acid fuchsin orange G stain, 400× magnification).

Fetal Outcome of 27 Pregnancies in Women with aHUS

The fetal outcomes of all 27 pregnancies are documented in Table 5. The median gestational week at birth was 38 (interquartile range, 33–40 weeks). Nineteen pregnancies (70%) resulted in a term live birth and there were two preterm live births. In addition, two preterm infants were stillborn, and four pregnancies resulted in spontaneous abortions before the 21st gestational week (C.2, C.4).

Table 5.

Outcomes of 27 pregnancies in 14 women with aHUS

Patient	Preg	Year	Outcome	Sex	Gestational Age at Birth, wk	Mode of delivery	Birth Weight, g	Birth Height, cm	Head Circumference at Birth, cm	Assessment
A.1	1	2009	live birth	female	40+0	vaginal delivery	3010	53	33	AGA
	2	2012	live birth	male	38+2	vaginal delivery	3340	50	33	AGA
A.2	1	2010	live birth	male	37+3	vaginal delivery	3110	51	33	AGA
	2	2011	live birth	male	40+3	vaginal delivery	3956	52	36	AGA
	3	2014	stillbirth	male	33+3	vaginal delivery	2230	46	35	AGA
B.1	1	2004	live birth	female	38+0	C-section	2900	40	33	AGA
B.2	1	2003	live birth	female	41+3	C-section	4090	52	35	AGA
	2	2005	live birth	male	40+2	C-section	3560	50	36	AGA
B.3	1	2001	live birtha	unknown	unknown	C-section	unknown	unknown	unknown	unknown
	2	2006	live birtha	unknown	unknown	C-section	unknown	unknown	unknown	unknown
B.4	1	2005	live birth	male	38+3	C-section	3665	52	36.5	AGA
B.5	1	1981	live birth	male	40+?	emergency C-section	3270	unknown	unknown	AGA
	2	1989	live birth	female	40+5	vaginal delivery	3500	50	34	AGA
C.1	1	2011	live birth	female	40+6	emergency C-section	4450	53	38	LGA
	2	2014	live birth	female	37+3	C-section	3130	49	35	AGAb
	3	2015	live birth	female	38+0	C-section	3150	51	33.5	AGA
C.2	1	1993	fetal death/abortion	na	18	vaginal delivery	na	na	na	na
	2	1999	spontaneous abortion	na	7	vaginal delivery	na	na	na	na
	3	2000	spontaneous abortion	na	9+6	vaginal delivery	na	na	na	na
C.3	1	2002	live birth/premature	female	26+?	emergency C-section	888	unknown	unknown	AGAc
C.4	1	2012	spontaneous abortion	na	18+3	vaginal delivery	na	na	na	na
C.5	1	2006	stillbirth	male	31+1	emergency C-section	1210	na	na	SGA
C.6	1	2002	live birth	male	39+0	C-section	3760	53	37	AGA
	2	2013	live birth	male	38+3	C-section	3030	51	33	AGA
	3	2015	live birth	female	38+0	C-section	3240	49	34	AGA
C.7	1	2002	live birth	male	40+3	vaginal delivery	unknown	unknown	unknown	unknown
	2	2009	live birth	male	35+?	vaginal delivery	unknown	unknown	unknown	unknown

Preg, pregnancy; AGA, appropriate for gestational age; C-section, caesarian section; ?, unknown; LGA, large for gestational age; na, not applicable; SGA, small for gestational age.

^aLost to follow-up.

^bIntermediate care unit stay.

^cIntensive care unit stay.

Seven pregnancies were complicated by p-aHUS: two were full term live births (C.1, C.6); three were preterm births, two live births (C.3, C.7) and one stillbirth (C.5); and there were two early spontaneous abortions due to fetal death (C.2, C.4).

In total there were 21 live infants: 16 with birth weights appropriate for gestational age, and one who was large for gestational age; the weights of the remaining four infants are unknown. The overall rate of adverse fetal outcomes in the aHUS cohort was 22 per 100 pregnancies (95% CI, 10 to 41) and was 70 per 100 for those whose pregnancies were complicated by p-aHUS (95% CI, 33 to 93).

Renal Outcome and Pregnancy in aHUS

At the end of the study in June of 2015, the renal status of the 14 women in our aHUS pregnancy cohort was as follows: six have native kidneys (normal kidney function, two; CKD-G1, two (one had had p-aHUS); and CKD-G2 and CKD-G4, one each), whereas the remaining eight progressed to ESRD (caused by p-aHUS in one). Seven of these received renal allografts, of which three have stable renal function with CKD-G1T to -G3T, two have CKD-G4T, and two developed ESRD (caused by p-aHUS in one) (Table 1). The rate of ESRD in all 14 women from first pregnancy until 2015 was 6.36 per 100 person-years (95% CI, 2.56 to 13.11; Table 1).

Discussion

Pregnancy is a critical condition for women with aHUS because of its well documented ability to trigger episodes of TMA with the attendant risk of irreversible renal failure. These risks led to many women being counseled not to have children and to a considerable number deciding to remain intentionally childless.^13,14 Recently this approach has been criticized for being over-pessimistic,⁴ and this view is strongly re-enforced by our analysis of pregnancy in women with aHUS in the Vienna TMA cohort. Specifically, our results show that: (1) only a quarter of the 27 pregnancies studied were complicated by episodes of p-aHUS; (2) episodes of p-aHUS were three times more common in the retrospectively studied pregnancies (33%), which were managed like conventional low-risk pregnancies, compared with those who were studied prospectively (11%) in a specialized unit and in most cases treated with prophylactic FFP; and (3) fetal outcomes were excellent in those without antenatal episodes of p-aHUS. Our data show that the outcome of pregnancies in women with aHUS is much better than commonly appreciated.

The overall prevalence of p-aHUS was 26% in our cohort, with five out of seven presenting during the first pregnancy. Four were “classic” p-aHUS episodes presenting postpartum (three with an additional trigger) and three patients with an initial diagnosis of preeclampsia/HELLP. With an incidence of 3%–5%, preeclampsia is the most frequent complication after 20th gestational week,¹⁵ and its early clinical presentation can mimic p-aHUS, HELLP syndrome, thrombotic thrombocytopenic purpura, acute fatty liver of pregnancy, and SLE. Distinguishing between them can be challenging,^4,16 but the most discriminatory laboratory tests have been summarized.¹⁷ Glomerular endotheliosis is the morphologic hallmark of preeclampsia,¹⁸ whereas thrombosis within vessels and glomeruli is the central feature in aHUS,¹⁹ and HELLP syndrome leads to acute tubular necrosis.^20,21 The physiologic changes during pregnancy complicate interpretation of hematologic tests,²² proteinuria,²³ and serum concentrations of complement factors, because the reference ranges established for nonpregnant subjects are inappropriate during pregnancy.^24,25

In a large, multicenter study of women with p-aHUS, Bruel et al.²⁶ reported that 58% of episodes occurred during first pregnancy, whereas in our study 71% of episodes of p-aHUS occurred in the first pregnancy. However, both studies document that a previously uncomplicated pregnancy does not reduce the probability of developing p-aHUS in subsequent pregnancies. Similarly, neither the presence nor the nature of complement gene mutations appeared to influence the risk of p-HUS in either cohort. Currently, there are no factors that can be used to predict the risk of developing p-aHUS in a given pregnancy, other than the presence of one of the disease triggers identified in nonpregnant patients with aHUS. Caesarian section was recently proposed as a particular risk factor for p-aHUS.¹² Although the proportion of caesarian sections in our cohort was more than twice as high (65% for deliveries ≥21 weeks of gestation) compared with the general population,²⁷ only three out of 15 (20%) caesarean sections were associated with postpartum p-aHUS, and all three had additional potential triggers (bleeding, infection, stillbirth).

The goal for the management of pregnancy in aHUS is to minimize the risk of an episode of TMA and mitigate the injury caused should one develop. Two approaches have been suggested: prophylactic infusions of FFP,⁶ perhaps coupled with plasma exchange as recommended for patients with thrombotic thrombocytopenic purpura²⁸; or alternatively expectant management with immediate treatment with eculizumab should an episode of TMA occur. This is feasible because the safety of eculizumab in pregnancy has been established in patients with paroxysmal nocturnal hemoglobinuria,^29,30 and in a small number of individuals with p-aHUS.^31–38

In the absence of data from randomized, controlled trials, we have adopted a pragmatic approach and have increased the maintenance dose of prophylactic FFP for those already receiving it,³⁹ and have introduced it for those who were not. Our standard regimen is to infuse FFP (10–20 ml/kg body wt every other week⁴⁰), starting at 15th gestational week, and continue at the original dose for at least 12 weeks postpartum. Additionally, we have infused FFP on the day of delivery and again 1 day later. None of the six pregnancies in patients treated in this way were complicated by p-aHUS and none were accompanied by evidence of dysregulated complement activity. In addition, there were three pregnancies in which we did not use prophylactic FFP or increase maintenance dose; one of these was associated with p-aHUS and the two others were not. Accordingly, the prophylactic use of FFP during pregnancy in women with aHUS appears safe, but effectiveness to prevent p-aHUS needs yet to be determined.

Despite good outcome of pregnancies, the long-term renal prognosis was poor. However, it is impossible to know whether pregnancy influenced this prognosis or whether it simply reflects the natural history of aHUS more generally. The latter would be consistent with data from a recent international study in which women with complement gene mutations progressed more frequently to dialysis-dependent renal failure.²⁶

In conclusion, our study emphasizes the frequency of successful pregnancies in women with aHUS. Close monitoring of such pregnancies for episodes of TMA is essential, but currently, the available data do not allow firm recommendations about the best strategy to prevent them.

Concise Methods

Patients

Patients presenting with laboratory signs of TMA or/and biopsy-proven TMA in renal tissue are eligible to be enrolled in the Vienna TMA cohort of the Division of Nephrology and Dialysis at the Medical University of Vienna. Our study includes all enrolled female patients with at least one reported pregnancy until 2015. All patients enrolled in the study gave written informed consent, investigations are in accordance with the Declaration of Helsinki, and the ethics committee of the Medical University of Vienna approved the study. Subjects were classified as aHUS in the case of: laboratory signs of TMA, sequence variations within complement regulatory protein genes, and absence of secondary forms of TMA.⁴¹ Disease definitions, classifications, and calculations are detailed in the Supplemental Material. Data collection was on the basis of chart review. With respect to pregnancies, data were obtained by personal interview, review of hospital charts, and the Austrian mother/child book.

Laboratory Methods

Laboratory work-up was carried out at the Department of Laboratory Medicine, Medical University of Vienna, and at the 3rd Department of Internal Medicine, Research Laboratory, Semmelweis University in Budapest. Genetic investigations were performed according to standard procedures and as detailed in the supplement (www.kimcl.at, http://semmelweis.hu/kutlab/en/introduction).

Disclosures

None.