Postpartum acute kidney injury: a review of 99 cases

Abstract

Postpartum acute kidney injury (PPAKI) constitutes an important cause of obstetric AKI. It is associated with high maternal and fetal mortality in developing nations. The aim of this study is to survey the etiology and outcomes of PPAKI in a tertiary care Indian hospital. Ninety-nine patients, without prior comorbidities, treated for PPAKI, between 2005–2014 at M.S. Ramaiah Medical College, were included for analysis in this retrospective, observational study. AKI was analyzed in terms of maximal stage of renal injury attained as per RIFLE criteria. Outcomes included requirement for renal replacement therapy (RRT), maternal and fetal outcomes. PPAKI constituted 60% of all obstetric AKI cases. Median maternal age was 23 years and 52% of patients were primigravidas. Mean serum creatinine was 4.1 mg/dL. Failure (33%) and injury (31%) were the major categories as per RIFLE criteria. Thirty-nine percent of cases required RRT. Sepsis, particularly puerperal sepsis, was the leading causes of PPAKI (75% of cases) and maternal mortality (94% of deaths). Maternal and fetal mortality were 19% and 22% respectively. The incidence of cortical necrosis was 10.3%. Three patients required long-term RRT. In conclusion, consistent with other Indian literature, we report a high incidence of PPAKI. We found incremental mortality on moving from “Risk” to “Failure” category of RIFLE. PPAKI was associated with high maternal and fetal mortality with sepsis being the leading cause. Our study highlights the need for provision of better quality of maternal care and fetal monitoring to decrease mortality associated with PPAKI in developing countries.

Keywords:

• Obstetric renal failure
• postpartum acute kidney injury
• postpartum renal failure
• RIFLE criteria
• pregnancy-related acute kidney injury

Introduction

Pregnancy-related acute kidney injury (PRAKI) is a major cause of maternal and fetal morbidity and mortality in developing countries (references). With improvement in antenatal and postnatal care, the incidence of PRAKI in India has steadily declined from 22% in 1960s, to 9% in 1980s¹, and further down to 3–7% in 2000 s.^2,3 Postpartum acute kidney injury (PPAKI) constitutes 28%–72%^4–6 of PRAKI cases. While sepsis and postpartum hemorrhage (PPH) are the major etiologic factors for PPAKI in developing countries,^4,6 severe preeclampsia, PPH and HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome contribute to PPAKI in developed nations.^7,8 Rare causes include hemolytic uremic syndrome (HUS), which is an important cause of end-stage renal disease (ESRD) in this population.

We present our experience of managing over a period of 10 years from 2005–2014 with the aim of identifying risk factors and clinical outcomes of PPAKI.

Material and methods

In this prospective, observational study, 165 pregnant patients with a diagnosis of AKI were admitted to our institution from 2005–2014. Of these patients, 99 cases with PRAKI were identified as postpartum in onset. M.S. Ramaiah hospitals, attached to M.S Ramaiah Medical College, are a major tertiary care referral center, catering to patient population from North Bangalore, state of Karnataka and adjacent states. Records were analyzed for demographic characteristics, obstetric history and clinical profile on admission. Obstetric history included parity, history of antenatal follow up, location, nature of delivery, pregnancy, maternal and fetal outcomes.

Patients with preexisting diabetes mellitus, hypertension, chronic kidney disease defined as s.cr >1.5 mg/dL or the presence of proteinuria >1 + on dipstick or renal transplant recipients, contracted kidneys on ultrasound were excluded from analysis.

As it is a retrospective study, we have not obtained signed consent form from the patients and in our institute ethical committee clearance is sought for only prospective studies.

Definitions

PPAKI was defined as AKI diagnosed from the time of childbirth to six weeks postdelivery. AKI was defined on the basis of RIFLE (risk, injury, failure, loss of function and end-stage renal disease) criteria using serum creatinine (s.cr) and urine output (u/o) based cut offs.⁹ Studies utilizing RIFLE criteria have good correlation between RIFLE criteria and in-hospital mortality,¹⁰ even in obstetric population.¹¹ Preeclampsia was defined as blood pressure reading > 140/90 mmHg diagnosed for the first time after 20 weeks of gestation with 2 + proteinuria on dipstick. Severe preeclampsia was defined as the association of severe arterial hypertension (systolic arterial pressure (SAP) exceeding 160 mmHg and diastolic arterial pressure (DAP) exceeding 110 mmHg) or proteinuria ≥5 g/L or ≥ 3 + or signs of visceral involvement (headaches, visual disturbances, epigastric or right upper-quadrant pain). Eclampsia was defined as the presence of new-onset grand mal seizures in a woman with preeclampsia.¹² HELLP syndrome was defined by the combination of thrombocytopenia (<100 g/L), elevated liver enzymes (AST >70 UI/L) and hemolysis. Sepsis was defined as per criteria laid down by American College of Chest Physicians.¹³ Puerperal sepsis was defined as per the World Health Organization definition as any bacterial infection of the genital tract that occurs after the birth of a baby.¹⁴ It is usually more than 24 h after delivery before the symptoms and signs appear. If, however, the woman has had prolonged rupture of membranes or a prolonged labor without prophylactic antibiotics, then the disease may become evident earlier. In this context, women presenting with intrauterine death, ruptured membranes and evidence of sepsis was classified under puerperal sepsis. Post-partum hemorrhage was defined as a blood loss of ≥500 mL after vaginal delivery or ≥1000 mL after cesarean delivery or as noted in the medical record by a care provider¹⁴.

Outcomes

Patient outcomes were noted for mortality and renal outcomes were defined as recovery of renal function or requirement of renal replacement therapy (RRT), as well as the presence or absence of proteinuria. Renal recovery was defined as a decline in serum creatinine to ≤1.0 mg/dL and the presence of 24-h urine protein (UP) of <150 mg within six weeks of diagnosis of AKI. Patients who did not satisfy the criteria for renal recovery were subjected to a renal biopsy. The biopsy analyzed for light microscopy (H and E stain) and immunofluorescence to stain for Immunoglobulin IgG, A and M as well complement factors C3 and C1q as well as kappa and lambda light chains. Pregnancy outcomes (preterm, term, miscarriages, abortions and mortality) and fetal outcomes by (birth weight and mortality) were noted as well.

Statistical analysis

Data were analyzed using SPSS software version 18.0. Chi-square test was used to find the factors associated with the outcome. Statistical significance was tested at p < 0.05. Risk estimate was calculated and its 95% CI was analyzed. Results are given as number, mean, median and interquartile range for quantitative variables and percentages for nominal variables.

Results

PPAKI constituted 60% of PRAKI cases and 0.9% of all AKI cases in our institution. Table 1 outlines the epidemiologic data of this cohort. Median patient age was 23 years (interquartile range, IQR: 23–28 weeks. Fifty-two (52%) were primigravidas. Ninety-one patients (92%) reported receiving antenatal care while 89 patients (90%) underwent institutional deliveries. Seventy-three patients (n = 73) were referred from other institutions. Sixty patients (61%) had vaginal delivery, while 34 patients (34%) required cesarean section. The remaining (n = 5) required instrumentation for delivery.

Table 1. Postpartum AKI (n = 99)

Parameters	N	%^a
Age (years)
Median (IQR) – 23.0 (23.0, 28.0)
15–20	12	12
21–25	40	40
26–30	30	30
31–35	17	17
Order of pregnancy
Primigravida	52	52
Multigravida	47	48
Antenatal Care
Received	91	92
Did not receive	8	8
Institutional Delivery/Abortion	89	90

Note: ^aPercentages have been rounded off to the nearest two digit number.

All patients required admission into intensive care unit (ICU) with median ICU stay of 8 days (IQR 3.0, 13.0). Nine patients (10%) required >15 days in ICU. The average time to presentation from the day of delivery was 1.1 ± 0.7 days. The IQR of serum creatinine on presentation was 3.2 (2.3, 6.5). The data were checked for normality using Shapiro test and it was found not to be normal so the descriptive Serum creatinine was given in terms of IQR. In terms of maximal RIFLE category achieved, 33% cases were classified as “Failure” (n = 33), closely followed by “Injury” in 31% cases (n = 31) and “Risk” is 23% cases (n = 23). Only 10% and 3% of cases were classified as “Loss” and “ESRD,” respectively, on long-term follow-up.

Sixty patients (61%) cases of PPAKI were managed conservatively, while the remaining 39 patients (39%) required renal replacement therapy (AKI-RRT). The chief modality of RRT was intermittent hemodialysis (HD) or slow low efficiency dialysis (SLED) in 29 patients (74% of AKI-RRT). Continuous renal replacement therapy (CRRT) and peritoneal dialysis (PD) was performed in four and six patients, respectively (Table 2).

Table 2. Renal parameters during hospitalization (n = 99).

Parameters	Value
Mean serum creatinine (mg/dl)	4.1
RIFLE category n(%)
Risk	23 (23%)
Injury	31 (31%)
Failure	33 (33%)
Loss	09 (10%)
ESRD	03 (3%)
Conservative treatment n(%)	60 (61%)
Renal replacement therapy n(%)	39 (39%)
HD/SLED	29
PD	6
CRRT	4

The etiology of PPAKI was found to be multifactorial. Sepsis was noted in 74 patients (75%) and disseminated intravascular coagulation (DIC) in 39 patients (39%). Majority of sepsis was related to puerperal sepsis (n = 60, 61%). Fourteen patients (14%) had retained products of conception. Rare causes of sepsis included swine flu and mucormycosis. AKI was attributed to PPH in 12 patients (19%). Sixty-six patients (67%) required packed red blood cell (PRBC) transfusion, of which 13(13%) required greater than 10 units PRBC. Median PRBC requirement was three units IQR (2.5–8.5). Other rare causes of PPAKI included systemic lupus erythematosus (SLE), HUS and transfusion reaction (one each) (Table 3).

Table 3. Causes of AKI (multifactorial).

Parameters	Number
Sepsis	74 (75%)
Puerperal sepsis	60
Urinary tract infections	6
Mastitis, pneumonia	2 each
Septic thrombophlebiitis	1
Necrotizing fascitis	1
H1N1	1
Mucormycosis	1
DIC	39 (39%)
PPH	12 (12%)
Preeclampsia with HELLP	08 (8%)
SLE	01 (1%)
Transfusion reactions	01 (1%)
HUS	01 (1%)
Isolated postpartum PIH	01 (1%)

Data on outcomes were available on 95 patients (97%) because four patients were discharged against medical advice (Table 4). Maternal mortality was 19% (n = 18) of which sepsis with multiorgan failure contributed to 94% deaths (n = 17). Puerperal sepsis accounted for 15 deaths (83%). One death was attributed to PPH. Mortality in AKI-RRT group was 31% (n = 12/39), while mortality in the conservative management group was 10% (n = 6/60) (Table 5). In terms of fetal outcomes, 22% fetal mortality was noted. Fifty-seven neonates (58%) were born premature (<37 weeks of gestation). Within RIFLE classification, highest mortality was noted in ‘Failure’ category (24%) with incremental mortality on progressing from ‘Risk’ to ‘Failure’. Complete renal recovery was noted in 69 patients (70% of overall cohort) during the follow-up period. Of the remaining, five patients had partial recovery of renal function and 3 required long-term renal replacement therapy. Renal biopsies in these eight patients revealed partial cortical necrosis in five and complete cortical necrosis in three. The overall incidence of cortical necrosis was 10.3%.

Table 4. Patient and renal outcomes* (n = 95).

Parameters	Number
Maternal mortality	18 (19%)
Sepsis	17
PPH	1
Mortality in AKI-conservative management	6
Mortality in AKI-RRT	12
Mortality as per RIFLE category:  n(% of n within each category)
Risk	03 (13%)
Injury	05 (16%)
Failure	8 (24%)
Loss	2 (22%)
ESRD	0 (0%)
Fetal mortality	14 (22%)
Death-censored renal outcomes (n = 77)
Complete recovery	69 (86%)
Partial recovery^a	5
ESRD^a	3

Notes: *Four patients left against medical advise.

a Partial cortical necrosis: 5, Complete cortical necrosis: 3.

Table 5. Statistical analysis of RRT versus conservative management.

	Yes	No	Total
RRT	12	27	39
Conservative management	6	54	60
Chi = 6.85, p = 0.0088.

Discussion

PPAKI represented approximately 60% of overall PRAKI in our series. Studies from Indian subcontinent including those by Arora et al., Gopalani et al. and Hassan et al. reveal that PPAKI contributes to 70–83% of PRAKI.^4,6,15 Studies on PPAKI from developed nations are sporadic in nature and report a low incidence of PRAKI and PPAKI. Jonard et al., in a five-year retrospective study found the incidence of PPAKI to be 0.11%,⁷ when analyzed in relation to total number of pregnancies.

Sepsis, particularly puerperal sepsis, remained the leading cause of PPAKI in our study, similar to results from other studies from Indian subcontinent.^4,6,16 Lack of antenatal care and perinatal care are the likely explanations for puerperal sepsis in most series. In our study, the incidence of puerperal sepsis is surprisingly high considering a large majority of patients in our cohort received antenatal care and underwent institutional deliveries. This brings to focus the need for providing good quality maternal health care and scope for improvement in training provided in our primary and secondary health care centers, where majority of deliveries take place. This figure is in contrast with series from developed nations where reported incidence of maternal sepsis is 0%–11.3%.^7,8 In this context, it is interesting to note that more than half of our patients had antecedent pregnancy-induced hypertension (PIH), which resulted in intrauterine device (IUD) (approximately one-third of cases) and hence, puerperal sepsis in the setting of ruptured membranes, implicating the role of early identification of such patients and need for aggressive fetal monitoring. The incidence of PPH as an etiology for PPAKI was similar to that seen in series from India and Canada.^4,8 With respect to preeclampsia with HELLP syndrome, it is established that up to 30% of cases with HELLP can present in the postpartum period, similar to the 32% postpartum HELLP seen in our institution (data not shown) and are associated with abruptio placentae and DIC (17%). We noted only one patient with postpartum HUS.

Regarding severity of AKI and use of RIFLE criteria in our cohort, close to 43% of our cohort had significant renal failure as defined by ‘Failure’ and ‘Loss’ categories, while close to one-third of the cohort required RRT. Studies utilizing RIFLE criteria in PPAKI are limited. In a review of 69 cases with PPAKI in French ICUs, Jonard et al. reported similar findings, with 51% cases reported as ‘Failure’ category on the day of admission.⁷ While renal function recovered in 35 of 68 patients (51%) during the ICU stay, of the remaining with renal dysfunction at the end of the ICU stay (n = 33), 79% were classified as ‘Failure’ (n = 26). In this group of patients classified as ‘Failure,’ 18 patients (28% of overall cohort) required dialysis during hospitalization. Intuitively, patients categorized as ‘Risk’ category showed faster recovery of renal function than those categorized as ‘Failure.’ In our cohort, 3.8% of surviving patients (n = 3) required chronic HD, similar to 4% patients with chronic HD in study by Jonard et. al. The incidence of patients who require renal replacement therapy in other studies varied from 7% to 20%.^4,6,16 This variation is largely a result of greater severity of renal failure and high incidence of cortical necrosis seen in most studies from Indian subcontinent, which portend poor prognosis in terms of renal recovery. While most studies, including ours, have not focused on long-term follow-up of renal function, study by Jonard et al. is remarkable in highlighting that renal function in PRAKI may take up to 6–12 months to recover.⁷

We observed high maternal mortality (19%) and fetal mortality (22%) in our study, consistent with reported literature from Indian subcontinent, while western literature reports 0%–2.9%.^7,8 The variation is likely the result of higher incidence of sepsis in developing countries, which remains a main cause of maternal mortality, in contrast to higher incidence of PIH and PPH in developed nations along with favorable socioeconomic factors, which have led to low maternal mortality. AKI requiring RRT associated with significantly high mortality as compared to the group requiring conservative management.^17,18

Our study represents one of the largest series focusing on factors determining patient and short-term renal outcomes in postpartum AKI. However, we were unable to study the impact of medical interventions on the outcomes in our cohort.

Conclusion

The development of AKI in pregnancy is a major clinical challenge because it is necessary to consider the outcome of mother and fetus. It can be caused by specific pregnancy related diseases which are not fully understood. It is essential to focus on the periodic evaluation to improve maternal and perinatal outcomes. Despite improvements in antenatal care, our health care policies need to address the gaps in providing good quality pregnancy care which can go a long way in reducing maternal and fetal mortality.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.