Preventing Malaria in HIV-Infected Pregnant Women
Preventing Malaria in HIV-Infected Pregnant Women
During the 2-year enrollment period (December 2009 to December 2011), 533 women were screened, among whom the majority was eligible for the CNM trial, that is, with CD4 count >350. Therefore, two-thirds of the 432 women randomized (N = 292) were allocated to the CM trial, and one-third (N = 140) was allocated to the CNM trial (Fig. 1).
The Supplemental Digital Content (see Table, SDC 2, http://links.lww.com/QAI/A485) shows baseline characteristics of the randomized population. HIV-related characteristics differed between trials, but within each trial, populations were similar in the 2 treatment groups. In both trials, 6% of women had positive P. falciparum microscopy and 38% were positive by PCR detection. Sixty-three percent of women reported having slept under bed net the 3 previous nights.
After randomization, 1 woman in the CNM trial refused to continue before receiving the allocated MQ treatment. There were 2 protocol violations at enrollment (contraindication not reported) in the MQ group of the CM trial. The first woman experienced seizures 8 days after MQ intake and recovered after anticonvulsant treatment. She later reported history of epilepsy. The other reported history of depression at the second visit while she was well tolerating MQ. Both women were asked to discontinue MQ.
Seven (5.0%) and 19 (6.5%) women were lost to follow-up in the CNM and CM trials, respectively, 2 (1.4%) and 7 (2.4%) experienced spontaneous abortion, and 1 died of HIV-related disease (in the CM trial). Finally, 130/140 (92%) and 263/292 (90%) of women in the CNM and CM trials, respectively, were followed until delivery of a live or dead infant after 28 weeks of pregnancy (usual definition for stillbirth as opposed to spontaneous abortion), and 108 (77%) and 233 (80%), respectively, had the primary end point assessed.
The proportions of women not included in the mITT population were similar in different trials and treatment groups (Fig.1). Overall, women with missing primary end point had been more frequently diagnosed for HIV during the ongoing pregnancy (56% vs. 38%, P = 0.003), and their HIV diagnosis was more recent (2.4 vs. 19.0 months ago in median, P = 0.005). Other characteristics did not differ.
In both trials, the mean gestational age was 21.8 (SD: 3.8) weeks at the first visit, 27.9 weeks (SD: 2.9) at the second visit, and 33.3 weeks (SD: 2.2) at the third visit. For patients allocated to MQ, the mean dose administered at the first visit was 14.2 mg/kg (SD: 1.1 mg/kg). Accounting for physiologic weight gain during pregnancy, it was 13.7 mg/kg (SD: 1.0) at the second visit and 13.4 mg/kg (SD: 1.2) at the third visit. For 3 women, MQ intake was followed by early vomiting justifying second administration. Adherence to MQ-IPTp was total, except for 1 woman who left the clinic before the supervised intake. Ninety percent and 95% of women self-reported total adherence to CTX prophylaxis and ART during their whole follow-up, respectively.
Deviations in the administration of the allocated treatment concerned 12 women (11% of the mITT population) and 8 women (3%) in the CNM and CM trials, respectively (). Adverse drug reactions (ADRs) caused MQ discontinuation in 1 and 3 women of the CNM and CM trials, respectively. CTX was discontinued in 1 woman of the CM trial.
Two women had a positive placental blood smear, 1 in the CTX-alone group of each trial. The noninferiority results are shown in Table 1.
In the CM trial, CTX efficacy for the prevention of placental parasitemia was not more than 5% inferior to the association of CTX + MQ-IPTp. In the CNM trial, because of the small sample size obtained, noninferiority could not be conclusively assessed. However, high-calculated conditional power (95%, see SDC 1, http://links.lww.com/QAI/A485) indicated that the probability of detecting noninferiority was high, if we had continued enrollments to reach the target sample size.
Sensitivity analyses successively excluded, then counted as failures, the 9 women who had received quinine or Coartem during follow-up, all in the CM trial (see Table, SDC 3, http://links.lww.com/QAI/A485). These analyses corroborated results of the main analysis, although the second approach lacked sufficient power to reach statistical significance.
In the CM trial, using PCR for P. falciparum detection in the placental blood, we detected no infection (0/105) in women receiving CTX + MQ versus 5/103 infections with CTX alone (P = 0.03 and 0.06 in mITT and PP analyses, respectively; Table 2). The low incidence rate of malaria parasitemia did not differ between treatment groups. No statistically significant differences were either observed regarding peripheral parasitemia at delivery, maternal hemoglobin, birth weight, or prematurity in any trial.
Vomiting, nausea, dizziness, and fatigue were more frequently reported with MQ (36% vs. 0% in both trials for dizziness, 37% vs. 3% in the CNM trial, and 34% vs. 0% in the CM trial for vomiting, P < 0.0001; Table 3). The time of occurrence of these AEs ranged from few minutes (for early vomiting) to one day after intake and 90% were resolved within 3 days. Overall, 57% of women receiving MQ-experienced ADRs (122/213), mostly graded minor (62%) or moderate (33%). ADRs were more frequent after the first intake (see Figure, SDC 4, see http://links.lww.com/QAI/A485). Adverse neuropsychiatric reactions to MQ were not reported except for seizures in an erroneously enrolled epileptic woman.
Four cutaneous reactions were linked to CTX, among which 1 exanthema graded 3, caused treatment discontinuation. No SAE was linked to MQ or CTX by the DSMB except the seizures already mentioned.
Although not significantly, spontaneous abortions tended to be more frequent in the CTX + MQ treatment group of the CM trial (4.1% vs. 0.7%, P = 0.12) (see Table, SDC 5, http://links.lww.com/QAI/A485).
One polymalformative syndrome was associated with intrauterine fetal death, and one hydrocephaly required surgical drainage. Other congenital abnormalities were benign. Finally, 7 early neonatal deaths and 11 later infant deaths were recorded (1.8% and 2.8%, respectively), mostly because of neonatal infection and acute diarrhea. After reviewing all SAEs, the DSMB did not see strong evidence for a link with the study drugs.
Implementation of PMTCT in this study has been detailed elsewhere. Only 4/393 women delivered without receiving ART. The end-of-pregnancy viral load was undetectable for 159 of the 221 women tested (71.9%). Of the 388 live infants born, 378 (97.4%) received ART during the first weeks of life; 2/3 were breastfed for a median duration of 6 months and 1/3 received formula milk; only 1 received mixed feeding.
HIV status could be determined in 229 of the 329 infants followed for more than 6 weeks (70%). Two infants were HIV infected, both in the CM trial (1 in each treatment group). The MTCT rate was 0.9% (95% exact CI: 0.1% to 3.1%).
Results
Trials Profiles and Baseline Characteristics
During the 2-year enrollment period (December 2009 to December 2011), 533 women were screened, among whom the majority was eligible for the CNM trial, that is, with CD4 count >350. Therefore, two-thirds of the 432 women randomized (N = 292) were allocated to the CM trial, and one-third (N = 140) was allocated to the CNM trial (Fig. 1).
The Supplemental Digital Content (see Table, SDC 2, http://links.lww.com/QAI/A485) shows baseline characteristics of the randomized population. HIV-related characteristics differed between trials, but within each trial, populations were similar in the 2 treatment groups. In both trials, 6% of women had positive P. falciparum microscopy and 38% were positive by PCR detection. Sixty-three percent of women reported having slept under bed net the 3 previous nights.
Follow-up
After randomization, 1 woman in the CNM trial refused to continue before receiving the allocated MQ treatment. There were 2 protocol violations at enrollment (contraindication not reported) in the MQ group of the CM trial. The first woman experienced seizures 8 days after MQ intake and recovered after anticonvulsant treatment. She later reported history of epilepsy. The other reported history of depression at the second visit while she was well tolerating MQ. Both women were asked to discontinue MQ.
Seven (5.0%) and 19 (6.5%) women were lost to follow-up in the CNM and CM trials, respectively, 2 (1.4%) and 7 (2.4%) experienced spontaneous abortion, and 1 died of HIV-related disease (in the CM trial). Finally, 130/140 (92%) and 263/292 (90%) of women in the CNM and CM trials, respectively, were followed until delivery of a live or dead infant after 28 weeks of pregnancy (usual definition for stillbirth as opposed to spontaneous abortion), and 108 (77%) and 233 (80%), respectively, had the primary end point assessed.
The proportions of women not included in the mITT population were similar in different trials and treatment groups (Fig.1). Overall, women with missing primary end point had been more frequently diagnosed for HIV during the ongoing pregnancy (56% vs. 38%, P = 0.003), and their HIV diagnosis was more recent (2.4 vs. 19.0 months ago in median, P = 0.005). Other characteristics did not differ.
In both trials, the mean gestational age was 21.8 (SD: 3.8) weeks at the first visit, 27.9 weeks (SD: 2.9) at the second visit, and 33.3 weeks (SD: 2.2) at the third visit. For patients allocated to MQ, the mean dose administered at the first visit was 14.2 mg/kg (SD: 1.1 mg/kg). Accounting for physiologic weight gain during pregnancy, it was 13.7 mg/kg (SD: 1.0) at the second visit and 13.4 mg/kg (SD: 1.2) at the third visit. For 3 women, MQ intake was followed by early vomiting justifying second administration. Adherence to MQ-IPTp was total, except for 1 woman who left the clinic before the supervised intake. Ninety percent and 95% of women self-reported total adherence to CTX prophylaxis and ART during their whole follow-up, respectively.
Deviations in the administration of the allocated treatment concerned 12 women (11% of the mITT population) and 8 women (3%) in the CNM and CM trials, respectively (). Adverse drug reactions (ADRs) caused MQ discontinuation in 1 and 3 women of the CNM and CM trials, respectively. CTX was discontinued in 1 woman of the CM trial.
Primary End Point
Two women had a positive placental blood smear, 1 in the CTX-alone group of each trial. The noninferiority results are shown in Table 1.
In the CM trial, CTX efficacy for the prevention of placental parasitemia was not more than 5% inferior to the association of CTX + MQ-IPTp. In the CNM trial, because of the small sample size obtained, noninferiority could not be conclusively assessed. However, high-calculated conditional power (95%, see SDC 1, http://links.lww.com/QAI/A485) indicated that the probability of detecting noninferiority was high, if we had continued enrollments to reach the target sample size.
Sensitivity analyses successively excluded, then counted as failures, the 9 women who had received quinine or Coartem during follow-up, all in the CM trial (see Table, SDC 3, http://links.lww.com/QAI/A485). These analyses corroborated results of the main analysis, although the second approach lacked sufficient power to reach statistical significance.
Secondary End Points
In the CM trial, using PCR for P. falciparum detection in the placental blood, we detected no infection (0/105) in women receiving CTX + MQ versus 5/103 infections with CTX alone (P = 0.03 and 0.06 in mITT and PP analyses, respectively; Table 2). The low incidence rate of malaria parasitemia did not differ between treatment groups. No statistically significant differences were either observed regarding peripheral parasitemia at delivery, maternal hemoglobin, birth weight, or prematurity in any trial.
Adverse Events in the Mother
Vomiting, nausea, dizziness, and fatigue were more frequently reported with MQ (36% vs. 0% in both trials for dizziness, 37% vs. 3% in the CNM trial, and 34% vs. 0% in the CM trial for vomiting, P < 0.0001; Table 3). The time of occurrence of these AEs ranged from few minutes (for early vomiting) to one day after intake and 90% were resolved within 3 days. Overall, 57% of women receiving MQ-experienced ADRs (122/213), mostly graded minor (62%) or moderate (33%). ADRs were more frequent after the first intake (see Figure, SDC 4, see http://links.lww.com/QAI/A485). Adverse neuropsychiatric reactions to MQ were not reported except for seizures in an erroneously enrolled epileptic woman.
Four cutaneous reactions were linked to CTX, among which 1 exanthema graded 3, caused treatment discontinuation. No SAE was linked to MQ or CTX by the DSMB except the seizures already mentioned.
Adverse Events in the Offspring
Although not significantly, spontaneous abortions tended to be more frequent in the CTX + MQ treatment group of the CM trial (4.1% vs. 0.7%, P = 0.12) (see Table, SDC 5, http://links.lww.com/QAI/A485).
One polymalformative syndrome was associated with intrauterine fetal death, and one hydrocephaly required surgical drainage. Other congenital abnormalities were benign. Finally, 7 early neonatal deaths and 11 later infant deaths were recorded (1.8% and 2.8%, respectively), mostly because of neonatal infection and acute diarrhea. After reviewing all SAEs, the DSMB did not see strong evidence for a link with the study drugs.
Mother-to-child Transmission of HIV
Implementation of PMTCT in this study has been detailed elsewhere. Only 4/393 women delivered without receiving ART. The end-of-pregnancy viral load was undetectable for 159 of the 221 women tested (71.9%). Of the 388 live infants born, 378 (97.4%) received ART during the first weeks of life; 2/3 were breastfed for a median duration of 6 months and 1/3 received formula milk; only 1 received mixed feeding.
HIV status could be determined in 229 of the 329 infants followed for more than 6 weeks (70%). Two infants were HIV infected, both in the CM trial (1 in each treatment group). The MTCT rate was 0.9% (95% exact CI: 0.1% to 3.1%).